PreprintPDF Available

A multi-lab study of bilingual infants: Exploring the preference for infant-directed speech

Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

From the earliest months of life, infants prefer listening to and learn better from infant-directed speech (IDS) than adult-directed speech (ADS). Yet, IDS differs within communities, across languages, and across cultures, both in form and in prevalence. This large-scale, multi-site study used the diversity of bilingual infant experiences to explore the impact of different types of linguistic experience on infants’ IDS preference. As part of the multi-lab ManyBabies project, we compared lab-matched samples of 333 bilingual and 385 monolingual infants’ preference for North-American English IDS (cf. ManyBabies Consortium, in press (MB1)), tested in 17 labs in 7 countries. Those infants were tested in two age groups: 6–9 months (the younger sample) and 12–15 months (the older sample). We found that bilingual and monolingual infants both preferred IDS to ADS, and did not differ in terms of the overall magnitude of this preference. However, amongst bilingual infants who were acquiring North-American English (NAE) as a native language, greater exposure to NAE was associated with a stronger IDS preference, extending the previous finding from MB1 that monolinguals learning NAE as a native language showed a stronger preference than infants unexposed to NAE. Together, our findings indicate that IDS preference likely makes a similar contribution to monolingual and bilingual development, and that infants are exquisitely sensitive to the nature and frequency of different types of language input in their early environments.
Content may be subject to copyright.
A multi-lab study of bilingual infants: Exploring the preference for infant-directed speech1
Krista Byers-Heinlein1, Angeline Sin Mei Tsui2, Christina Bergmann3, Alexis Black4, Anna2
, Maria Julia Carbajal
, Samantha Durrant
, Christopher T. Fennell
, Anne-Caroline
Fiévet6, Michael C. Frank2, Anja Gampe8, Judit Gervain9, Nayeli Gonzalez-Gomez10, J.4
Kiley Hamlin4, Naomi Havron6, Mikołaj Hernik11, Shila Kerr12, Hilary Killam1, Kelsey5
Klassen13, Jessica Kosie14, Ágnes Melinda Kovács15, Casey Lew-Williams14, Liquan Liu16 ,6
Caterina Marino
, Meghan Mastroberardino
, Victoria Mateu
, Clarie Noble
, Adriel John
, Linda Polka
, Christine E. Potter
, Leher Singh
, Melanie Soderstrom
, Megha
Sundara17, Connor Waddell16, Janet Werker4, & Stephanie Wermelinger8
1Concordia University10
2Stanford University11
3Max Planck Institute for Psycholinguistics12
4University of British Columbia13
5University of Liverpool14
6ENS, EHESS, CNRS, PSL University15
7University of Ottawa16
8University of Zurich17
9Integrative Neuroscience and Cognition Center (INCC), CNRS & Université Paris18
10 Oxford Brookes University20
11 UiT The Arctic University of Norway21
12 McGill University, School of Communication Sciences and Disorders22
13 University of Manitoba23
14 Princeton University24
15 Central European University25
16 Western Sydney University26
17 UCLA27
18 National University of Singapore28
Author Note29
"Individual participating labs acknowledge funding support from: the Natural Sciences
and Engineering Research Council of Canada (402470-2011 and 2015-03967); the Social31
Sciences and Humanities Research Council of Canada Insight Grant (435-2015-1974 and32
435-2015-0385); Agence Nationale de la Recherche (ANR-17-EURE-0017 and33
ANR-10-IDEX-0001-02); Western Sydney University Early Career Researcher Start-up Grant
(20311.87608); European Commission (MSCA-IF-798658); a European Research Council35
Synergy Grant, SOMICS (609819); ERC Consolidator Grant "BabyRhythm" (773202); The36
Leverhulme Trust (ECF-2015-009); The UK Economic and Social Research Council37
(ES/L008955/1); Research Manitoba, Children’s Hospital Research Institute of Manitoba,38
University of Manitoba; ODPRT funds, National University fo Singapore; and the National
Institute of Child Health and Human Development (R01HD095912)."40
Correspondence concerning this article should be addressed to Krista Byers-Heinlein,41
7141 Sherbrooke St. West, Montreal, QC, H4B 1R6, Canada. E-mail:
From the earliest months of life, infants prefer listening to and learn better from44
infant-directed speech (IDS) than adult-directed speech (ADS). Yet, IDS differs within45
communities, across languages, and across cultures, both in form and in prevalence. This46
large-scale, multi-site study used the diversity of bilingual infant experiences to explore the47
impact of different types of linguistic experience on infants’ IDS preference. As part of the48
multi-lab ManyBabies project, we compared lab-matched samples of 333 bilingual and 38549
monolingual infants’ preference for North-American English IDS (cf. ManyBabies50
Consortium, in press (MB1)), tested in 17 labs in 7 countries. Those infants were tested in51
two age groups: 6–9 months (the younger sample) and 12–15 months (the older sample). We
found that bilingual and monolingual infants both preferred IDS to ADS, and did not differ
in terms of the overall magnitude of this preference. However, amongst bilingual infants who
were acquiring North-American English (NAE) as a native language, greater exposure to55
NAE was associated with a stronger IDS preference, extending the previous finding from56
MB1 that monolinguals learning NAE as a native language showed a stronger preference57
than infants unexposed to NAE. Together, our findings indicate that IDS preference likely58
makes a similar contribution to monolingual and bilingual development, and that infants are
exquisitely sensitive to the nature and frequency of different types of language input in their
early environments.61
Keywords: language acquisition; bilingualism; speech perception; infant-directed speech;
reproducibility; experimental methods63
Word count: 1305464
A multi-lab study of bilingual infants: Exploring the preference for infant-directed speech65
When caregivers interact with their infants, their speech often takes on specific,66
distinguishing features in a speech register known as infant-directed speech (IDS; Fernald et
al., 1989). IDS is produced by caregivers of most (although not all) linguistic and cultural68
backgrounds, and is typically characterized by a slow, melodic, high-pitched, and69
exaggerated cadence (Farran, Lee, Yoo, & Oller, 2016; Fernald et al., 1989; Kitamura,70
Thanavishuth, Burnham, & Luksaneeyanawin, 2001; Pye, 1986; Shute & Wheldall, 1999).71
From early in life, infants tune their attention to IDS, preferring to listen to IDS over72
adult-directed speech (ADS) both at birth (Cooper & Aslin, 1990), as well as later in infancy
(Cooper, Abraham, Berman, & Staska, 1997; Cooper & Aslin, 1994; Fernald, 1985; Hayashi,
Tamekawa, & Kiritani, 2001; Kitamura & Lam, 2009; Newman & Hussain, 2006; Pegg,75
Werker, & McLeod, 1992; Santesso, Schmidt, & Trainor, 2007; Singh, Morgan, & Best, 2002;
Werker & McLeod, 1989; Werker, Pegg, & McLeod, 1994).77
Infants’ preference for IDS may play a useful role in early language learning. For78
example, infants are better able to discriminate speech sounds in IDS than in ADS (Karzon,
1985; Trainor & Desjardins, 2002), more efficiently segment words from continuous speech in
an IDS register (Thiessen, Hill, & Saffran, 2005), demonstrate better long-term memory for
words spoken in IDS (Singh, Nestor, Parikh, & Yull, 2009) and learn new words more82
effectively from IDS than ADS (Graf Estes & Hurley, 2013; Ma, Golinkoff, Houston, &83
Hirsh-Pasek, 2011; but see Schreiner, Altvater-Mackensen, & Mani, 2016).84
While most studies have confirmed a general, early preference for IDS, to date there is
very little research aimed at understanding how different linguistic experiences affect infants’
preferences. For instance, although the existence of IDS has been demonstrated in a large87
number of cultures (see above citations), the vast majority of the research on infants’ IDS88
preferences has been conducted in North America, using English speech typically directed at
North American English-hearing infants (Dunst, Gorman, & Hamby, 2012). Most critically,90
past work has been limited to a particular kind of linguistic (and cultural) experience: that
of the monolingual infant. Here, we present a large-scale, multi-site, pre-registered study on
bilingual infants, a population that is particularly suited to explore the relationship between
language experience and IDS preference. Moreover, this research provides important insight
into the early development of bilingual infants, a large but understudied population.95
Does experience tune infants’ preference for IDS?96
What role might experience play in tuning infants’ attention to IDS? We aggregated97
results from a recent published meta-analysis (Dunst et al., 2012) with additional98
community-contributed data (MetaLab, 2017) to examine their combined results. When all
62 studies are considered, we found a moderately-sized average effect of Cohen’s d=.64. A100
focus on the 22 studies most similar to ours (testing IDS preference using looking times101
collected in a laboratory, among typically-developing infants from 3–15 months, with102
naturally-produced English-spoken IDS from an unfamiliar female speaker), the effect size is
slightly lower,
= .6. Although this meta-analysis focused on infants in the first year of life,
other studies of infants aged 18–21 months have also reported a preference for IDS over ADS
(Glenn & Cunningham, 1983; Robertson, von Hapsburg, & Hay, 2013). There is some106
evidence that older infants show a greater preference for IDS than younger infants (Dunst et
al., 2012), although an age effect was not found in the subsample of 22 studies mentioned108
above. More evidence is needed to explore the possibility that increased language experience
as children grow enhances their preference for IDS.110
Another variable that would be important in understanding the role of experience in111
the preference for IDS is whether the speech stimuli were presented in a native or non-native
language. Numerous studies in early perception find different developmental trajectories for
perception of native versus non-native stimuli (e.g. discriminating human faces114
vs. discriminating monkey faces, Lewkowicz & Ghazanfar, 2006; discriminating native115
vs. discriminating non-native speech sound categories, Maurer & Werker, 2014; segmenting116
word forms from fluent speech, e.g., Polka & Sundara, 2012). Generally, whereas infants117
show increasing proficiency in discriminating the types of faces and sounds that are present
in their environment, they lose sensitivity to the differences between non-native stimuli over
time. This general pattern might lead us to predict that infants will initially be sensitive to
differences between IDS and ADS in both the native and non-native languages, but that this
initial cross-linguistic sensitivity will decline with age. In other words, at some ages, infants’
preference for IDS over ADS could be enhanced when hearing their native language.123
However, to date, there is very little data on this question. Importantly, this general trend, if
it exists, may interact with differences across languages in the production of IDS. The125
exaggerated IDS of North American English might be either more interesting or less126
interesting to an infant whose native language is characterized by a less exaggerated form127
IDS, than for an infant who regularly hears North American English IDS.128
Only a handful of IDS preference studies have explicitly explored infants’ preference for
IDS from infants’ native versus a non-native language. Werker et al. (1994) compared 4.5-130
and 9-month-old English and Cantonese-learning infants’ preference for videos of Cantonese
mothers using IDS versus ADS. Both groups showed a preference for IDS; however, the132
magnitude of the preference between the two groups was not specifically compared (Werker
et al., 1994). Hayashi et al. (2001) studied Japanese-learning infants’ (aged 4–14 months)134
preference for native (Japanese) and non-native (English) speech. Japanese-learning infants
generally showed a preference for Japanese IDS over ADS, as well as an increasing preference
for Japanese IDS over English IDS. The latter finding shows that infants tune into their137
native language with increased experience; however, as the study did not measure infants’138
interest in English ADS, we do not know whether Japanese infants were equally sensitive to
the difference between ADS and IDS in the non-native stimuli, or whether/how this might140
change over time.141
Infants growing up bilingual are typically exposed to IDS in two languages. They142
provide a particularly useful wedge in understanding experiential influences on infants’143
attention to IDS. Bilingual infants receive less exposure to each of their languages than144
monolingual infants, and the exact proportion of exposure to each of their two languages145
varies from infant to infant. This divided exposure does not appear to slow the overall rate146
of language acquisition: bilinguals pass their language milestones on approximately the same
schedule as monolingual infants, such as the onset of babbling and the production of their148
first words (Werker & Byers-Heinlein, 2008). Nonetheless, children from different language149
backgrounds receive different types of input, and must ultimately acquire different language
forms, which can alter some patterns of language acquisition (e.g., Choi & Bowerman, 1991;
Slobin, 1985; Tardif, 1996; Tardif, Shatz, & Naigles, 1997; Werker & Tees, 1984). As a152
consequence, bilingual infants allow researchers to investigate how a given “dose” of153
experience with a specific language relates to phenomena in language acquisition, while154
holding infants’ age and total experience with language constant (Byers-Heinlein & Fennell,
Aside from the opportunity to study dose effects, it is important to examine the157
preference for IDS in bilingual infants for the sake of understanding bilingual development158
itself. Several lines of research suggest that early exposure to two languages changes some159
aspects of early development (Byers-Heinlein & Fennell, 2014), including bilinguals’160
perception of non-native speech sounds (i.e., sounds that are in neither of their native161
languages). For example, a number of studies have reported that bilinguals maintain162
sensitivity to non-native consonant contrasts (García-Sierra, Ramírez-Esparza, & Kuhl, 2016;
Petitto et al., 2012; Ramírez, Ramírez, Clarke, Taulu, & Kuhl, 2017), tone contrasts (Graf164
Estes & Hay, 2015; Liu & Kager, 2017a), and visual differences between languages (i.e.,165
rhythmic and phonetic information available on the face of talkers; Sebastián-Gallés,166
Albareda-Castellot, Weikum, & Werker, 2012) until a later age than monolinguals. Other167
studies have suggested that bilinguals’ early speech perception is linked to their language168
dominance (Liu & Kager, 2015; Molnar, Carreiras, & Gervain, 2016; Sebastián-Gallés &169
Bosch, 2002), whereby bilinguals’ perception most closely matches that of monolinguals in170
their dominant language. Bilingual infants also demonstrate some cognitive differences from
monolinguals that are not specific to language, including faster visual habituation (Singh et
al., 2015), better memory generalization (Brito & Barr, 2014; Brito, Sebastián-Gallés, &173
Barr, 2015), and greater cognitive flexibility (Kovács & Mehler, 2009a, 2009b). This might174
reflect an early-emerging difference in information processing between the two groups.175
Together, these lines of work raise the possibility that preference for IDS versus ADS could176
have a different developmental course for bilingual and monolingual infants, and that177
bilinguals’ distinct course could interact with factors such as language dominance.178
Bilinguals’ exposure to and learning from IDS179
Overall, there is very little research on whether bilinguals’ experience with IDS is180
comparable to monolinguals’ experience. Some research has compared English monolinguals
and English-Spanish bilinguals in the United States (Ramírez-Esparza, García-Sierra, &182
Kuhl, 2014, 2017). Here, researchers reported that bilingual infants around 1 year of age183
received less exposure to IDS than monolingual infants on average. Moreover, in the184
bilingual families, input was more evenly distributed across infant- and adult-directed185
registers. It is difficult to know whether the results reported in these studies generalize to186
other populations of bilinguals, or whether it was specific to this language community. As187
acknowledged by the authors, the bilinguals in this study were of a lower SES than the188
monolinguals, which could have driven differences in the amount of IDS that infants heard.189
On the other hand, it might be the case that bilingual infants more rapidly lose their190
preference for the IDS register than do monolinguals, and that caregivers of bilinguals191
respond to this by reducing the amount of IDS input they provide.192
Bilingual infants might also hear IDS that differs prosodically and phonetically from193
that heard by monolingual infants. Bilingual infants often have bilingual caregivers, and194
even when they are highly proficient speakers, their speech may vary from that of195
monolinguals. One study compared vowels produced in the IDS of monolingual English,196
monolingual French, and balanced French-English bilingual mothers living in Montreal197
(Danielson, Seidl, Onishi, Alamian, & Cristia, 2014). Bilingual mothers’ vowels were distinct
in the two languages, and the magnitude of the difference between French and English199
vowels was similar to that shown by monolingual mothers. However, another study showed200
that in a word-learning task, 17-month-old French-English bilinguals learned new words201
better from a bilingual speaker than a monolingual speaker, even though acoustic202
measurements did not reveal what dimension infants were attending to (Fennell &203
Byers-Heinlein, 2014; similar findings were found in Mattock, Polka, Rvachew, & Krehm,204
2010). Finally, a study of Spanish-Catalan bilingual mothers living in Barcelona found that
some mothers were more variable in their productions of a difficult Catalan vowel contrast206
than monolingual mothers (Bosch & Ramon-Casas, 2011). Thus, bilingual infants may not207
only differ in the amount of IDS they hear in a particular language relative to monolingual208
infants, but different populations of bilingual infants may also vary in how similar the IDS209
they hear is to monolingual-produced IDS in the same languages. This could, in turn, lead210
to greater variability across bilinguals in their preference for IDS over ADS when tested with
any particular stimulus materials.212
Regardless of bilingual infants’ specific experience with IDS, evidence suggests that213
bilinguals might enjoy the same learning benefits from IDS as monolinguals. For example,214
Ramírez-Esparza et al. (2017) found that greater exposure to IDS predicted larger215
vocabulary size in both monolingual and bilingual infants. Indeed, an untested possibility is
that exposure to IDS might be of particular benefit to bilingual infants. Bilinguals face a217
more complex learning situation than monolinguals, as they acquire two sets of sounds,218
words, and grammars simultaneously (Werker & Byers-Heinlein, 2008). This raises the219
possibility that bilingual infants might have enhanced interest in IDS relative to220
monolinguals, or that they might maintain a preference for IDS until a later age than221
monolinguals, similiar to the extended sensitivity observed in bilingual infants’ perception of
non-native phonetic contrasts.223
Replicability in research with bilingual infants224
Working with bilingual infant populations engenders unique replicability issues above225
and beyond those common in the wider field of infant research (e.g., between-lab variability,
methodological variation, etc.; see Frank et al., 2017). These issues begin with the nature of
the population. Our discussion of bilingual infants thus far has used “bilingual” as a blanket
term to describe infants growing up hearing two or more languages. However, this usage229
belies the large variability in groups of infants described as “bilingual”. First, some studies of
bilinguals have included infants from a homogeneous language background (where all infants
are exposed to the same language pair; e.g. English-Spanish in Ramírez-Esparza et al., 2017),
while others have included infants from heterogeneous language backgrounds (where infants
are exposed to different language pairs, e.g., English-Other, where “Other” might be Spanish,
French, Mandarin, Punjabi, etc.; e.g., Fennell, Byers-Heinlein, & Werker, 2007). Second,235
some bilinguals learn two typologically closely related languages (e.g. Spanish-Catalan) while
others learn two distant languages (e.g. English-Mandarin). Third, there is wide variability237
between bilingual infants in the amount of exposure to each language, which introduces an238
extra dimension of individual difference relative to studies with monolingual infants. Fourth,
studies define bilingualism in different ways, ranging from a liberal criterion of at least 10%
exposure to the non-dominant language to at least 40% exposure to the non-dominant241
language (Byers-Heinlein, 2015). Finally, bilingual and monolingual populations can be242
difficult to compare because of cultural, sociological, and socio-economic status differences243
that exist between samples.244
All of the above difficulties have resulted in very few findings being replicated across245
different samples of bilinguals. The limited research that has compared different types of246
bilingual learners has indicated that the particular language pair being learned by bilingual
infants influences speech perception of both native (Bialystok, Luk, & Kwan, 2005; Sundara
& Scutellaro, 2011) and non-native (Patihis, Oh, & Mogilner, 2015) sounds. In contrast,249
other studies have not found differences between bilinguals learning different language pairs,
for example in their ability to apply speech perception skills to a word learning task (Fennell
et al., 2007). Generally, we do not know how replicable most findings are across different252
groups of bilinguals, or how previously reported effects of bilingualism on learning and253
perception are impacted by the theoretically interesting moderators discussed above.254
Research on bilingual infants also faces many of the same general concerns shared with
other fields of infancy research, such as challenges recruiting sufficient participants to256
conduct well-powered studies (Frank et al., 2017). Finding an appropriate bilingual sample257
further limits the availability of research participants, even in locations with significant258
bilingual populations. Such issues are particularly relevant given the recent emphasis on the
replicability and best practices in psychological science (Klein et al., 2014; Open Science260
Collaboration, 2015; Simmons, Nelson, & Simonsohn, 2011). Of particular interest is261
whether bilingual infants as a group show greater variability in their responses than262
monolingual infants, and how to characterize the variability of responses between the263
different types of samples of bilinguals that can be recruited by particular labs (i.e.,264
homogeneous vs. heterogeneous samples). Understanding whether variability differs265
systematically across groups is vital for planning appropriately-powered studies.266
Description of the current study267
Here, we report a large-scale, multi-site, pre-registered study aimed at using data from
bilingual infants to understand variability in infants’ preference for IDS over ADS. This269
study, “ManyBabies 1 Bilingual”, is a companion project to the “ManyBabies 1” project,270
published in a previous issue of this journal (ManyBabies Consortium, in press). The two271
studies were conducted in parallel, using the same stimuli and experimental procedure.272
However, while ManyBabies 1 analyzed all data collected from monolingual infants (including
those data from monolinguals reported here), the current study reports additional data from
bilingual infants not reported in that paper. Our multi-site approach gives us precision in275
estimating the overall effect size of bilingual infants’ preference for IDS, while also allowing276
us to investigate how different types of language experience moderate this effect.277
Our primary approach was to compare bilinguals’ performance to the performance of278
monolinguals tested in the same lab, that is, a subset of the data reported in the279
ManyBabies 1 paper. This approach has two notable advantages. First, within each lab,280
bilinguals shared one of their two languages with monolinguals (the language of the wider281
community). Second, testing procedures were held constant within each lab. Thus, this282
approach allowed us to minimize procedural confounds with infants’ bilingual status.283
However, a disadvantage of this approach is that it leaves out data from monolingual infants
tested in other labs, which could potentially add precision to the measured effects. Thus, we
performed secondary analyses comparing all bilinguals to all monolinguals within the same286
age bins, regardless of the labs each had been tested in.287
Another important difference is in the age groups tested. The ManyBabies 1 study288
tested monolinguals in four equal age windows: 3–6 months, 6–9 months, 9–12 months, and
12–15 months. Due to limitations in the numbers of bilingual infants that could be recruited,
we tested bilinguals in only two of these age windows: 6–9 months, and 12–15 months. The
specific age bins selected were based on a preliminary survey of availability of the age ranges
from participating laboratories. The choice of non-adjacent age bins also increased the293
chances of observing developmental differences.294
All infants were tested using the same stimuli, which consisted of recordings of295
North-American English (NAE) accented IDS and ADS. Because of the international nature
of this multi-site project, these stimuli were native for some infants but non-native for other
infants, both in terms of the language of the stimuli (English), and the variety of298
infant-directed speech (NAE-IDS is particularly exaggerated in its IDS characteristics299
relative to other varieties of IDS; see Soderstrom, 2007 for a review). Moreover, the stimuli300
were produced by monolingual mothers. Thus, infants’ exposure to the type of stimuli used
varied from low (monolinguals and bilinguals not exposed to NAE), to moderate (bilinguals
learning NAE as one of their two languages), to high (monolinguals learning NAE).303
Infants were tested in one of three experimental setups regularly used to test infant304
auditory preference: central fixation, eye-tracking, and headturn preference procedure. The
use of a particular setup was the choice of each lab, depending on their equipment and306
expertise. Labs that tested both monolinguals and bilinguals used the same setup for both307
groups. On all setups, infants heard a series of trials presenting either IDS or ADS, and their
looking time to an unrelated visual stimulus (e.g., a checkerboard) was used as an index of309
their attention. In central fixation, infants sat in front of a single screen that displayed a310
visual stimulus, and their looking was coded via button press using a centrally positioned311
camera while the auditory stimulus played. Eye-tracking was similar, except that infants’312
looking was coded automatically using a corneal-reflection eye-tracker. In the headturn313
preference procedure setup (HPP; see Kemler Nelson et al., 1995), infants sat in the middle
of a room facing a central visual stimulus. Their attention was drawn to the left or right side
of the room by a visual stimulus while the auditory stimulus played, and the duration of316
their looking was measured via button press using a centrally positioned camera.317
Research questions318
We identified three basic research questions addressed by this study. Note that it was319
not always possible to make specific predictions given the very limited data on infants’320
cross-language preferences for IDS over ADS, and particularly the absence of data from321
bilingual infants. We also note that the ManyBabies 1 project, focusing on monolingual322
infants, addresses other more general questions such as the average magnitude of the IDS323
preference, changes in preference over age, and the effects of methodological variation324
(ManyBabies Consortium, in press). The main questions addressed by data from bilingual325
infants are:326
1. How does bilingualism affect infants’ interest in IDS relative to ADS? As described327
above, monolingual infants display an early preference for IDS that grows in strength328
at least through the first year of life. We anticipated that the bilingual experience329
might result in a different pattern of IDS preference; however, the direction and330
potential source of any difference is difficult to predict. For example, the more331
challenging nature of early bilingual environments might induce an even greater332
preference for IDS over ADS relative to monolinguals. This enhanced preference could
be shown across development, or might be observed only at certain ages. On the other
hand, given some evidence that parents of bilingual infants produce relatively less IDS
than parents of monolingual infants, it may be that bilinguals show less interest in IDS
than monolinguals. We also explored the following questions as potential sources for an
emerging difference between populations: If an overall difference between monolingual
and bilingual infants’ preference for IDS is observed, can this be accounted for by339
systematic differences in socioeconomic status? Do bilinguals show greater variability340
in their preference for IDS than monolinguals?341
2. How does the amount of exposure to NAE-IDS affect bilingual infants’ listening342
preferences? While we expected infants across different language backgrounds to show
greater interest in IDS over ADS, we investigated whether this was moderated by the344
amount of exposure to NAE. For monolinguals, this exposure would be either 100%345
(monolingual learners of NAE) or 0% (monolingual learners of other languages). For346
bilinguals, some infants would have 0% exposure to NAE-IDS (e.g., bilingual infants347
learning Spanish and Catalan) while others would have a range of different exposures348
(e.g., bilingual infants learning NAE and French). This allowed us to at least partially
disentangle dose effects of exposure to NAE-IDS from infants’ bilingualism. An350
additional possibility is that infants’ exposure to NAE would predict overall attention
to both infant-directed and adult-directed NAE, with no differential effects on interest
to IDS versus ADS. Finally, it is possible that NAE-IDS is equally engaging to infants
regardless of their experience with North American English.354
3. Finally, we had planned to ask how bilingual infants’ listening to NAE-IDS and ADS355
impacted by the particular language pair being learned. We intended to ask this356
question at both the group and at the individual level. At the group level, we planned
to investigate whether different patterns of results would be seen in homogeneous358
versus heterogeneous samples of bilinguals, in terms of overall preference for IDS and359
group-level variability. However, ultimately we had insufficient homogeneous samples360
to address this question. At the individual level, we were interested in how the361
particular language pair being learned modulated infants’ preference for IDS. As we362
did not know a priori what language pairs would have sufficient sample size for363
analysis, this was considered a potential exploratory analyses. Ultimately, due to the364
nature of our main results and the diverse language backgrounds of our final sample,365
we decided to leave this question open for future investigations.366
Our methods largely followed those used by the ManyBabies 1 monolingual companion
project (ManyBabies Consortium, in press), with the exception of the nature of the bilingual
participants tested. In this section, we will provide only a brief overview of the methods370
shared by the two studies, focusing specifically on areas where the two studies differ. We371
report how we determined our sample size, all data exclusions, all manipulations, and all372
measures in the study.373
Participation Details374
Time-frame. An open call for labs to participate was issued on February 2, 2017.375
Participant testing began on May 1, 2017. Testing for monolinguals ended on April 30, 2018.
Because of the additional difficulty of recruiting bilingual samples, the end-date for collection
of these data was extended by four months to August 31, 2018. Due to a miscommunication,
one lab continued testing data beyond this deadline but prior to data analysis, and these379
data were included in the final sample.380
Age distribution.
Labs contributing data from bilingual infants were asked to test
participants in at least one of two (but preferably both) age bins: 6–9 month-olds (6:1 – 9:0)
and 12–15 month-olds (12:1 – 15:0). Labs were asked to aim for a mean age at the centre of
the bin, with distribution across the entire age window.384
Lab participation criterion.. Considering the challenges associated with385
recruiting bilingual infants and the importance of counterbalancing in our experimental386
design, we asked labs to contribute a minimum of 16 infants per age and language group387
(note that infants who met inclusion criteria for age and language exposure but were388
ultimately excluded for other reasons counted towards this minimum N). We also expected389
that requiring a relatively low minimum number of infants would encourage more labs to390
contribute a bilingual sample, and under our statistical approach a larger number of groups
is more important than a larger number of individuals (Maas & Hox, 2005). However, labs392
were encouraged to contribute additional data provided that decisions about when to stop393
data collection were made ahead of time (e.g., by declaring an intended start and end date394
before data collection). A sensitivity analysis showed that, with a sample of 16 infants and395
assuming the average effect size of similar previous studies (Cohen’s d = .7; Dunst et al.,396
2012; MetaLab, 2017), individual labs would have 74% power to detect a preference for IDS
in a paired-samples t-test (alpha = .05, one-tailed). We note that some labs were unable to
recruit their planned minimum sample of 16 bilingual infants that met our inclusion criteria
in the timeframe available, a point we will return to later in the paper.400
Labs were asked to screen infants ahead of time for inclusion criteria, typically by401
briefly asking about language exposure over the phone. Despite this screening process, some
infants who arrived in the lab for testing fell between the criteria for monolingual and403
bilingual status based on the comprehensive questionnaire. In such cases, the decision404
whether to test the infant was left up to individual laboratories’ policy, but we asked that405
data from any babies who entered the testing room be submitted for data processing (even406
though some such data might be excluded from the main analyses).407
Each lab followed the ethical guidelines and ethics review board protocols of
their own institution. Labs submitted anonymized data for central analysis that identified409
participants by code only. Video recordings of individual participants were coded and stored
locally at each lab, and where possible were uploaded to a central controlled-access databank
accessible to other researchers.412
Defining bilingualism. Infants are typically categorized as bilingual as a function414
of their parent-reported relative exposure to their languages. However, studies vary415
considerably in terms of inclusion criteria for the minimum exposure to the non-dominant416
language, which in previous studies has ranged from 10% to 40% of infants’ exposure417
(Byers-Heinlein, 2015). Some bilingual infants may also have some exposure to a third or418
fourth additional languages. Finally, infants can vary in terms of when the onset of exposure
to their additional languages is, which can be as early as birth or anytime thereafter. We420
aimed to take a middle-of-the-road approach to defining bilingualism, attempting to balance
a need for experimental power with interpretable data.422
Thus, we asked each participating lab to recruit a group of simultaneous bilingual423
infants who were exposed to two languages between 25% and 75% of the time, with regular
exposure to both languages beginning within the first month of life. There was no restriction
as to whether infants were exposed to additional languages, thus some infants could be426
considered multilingual (although we continue to use the term bilingual throughout this427
manuscript). These criteria would include, for example, an infant with 40% English, 40%428
French, and 20% Spanish exposure, but would exclude an infant with 20% English, 70%429
French, and 10% Spanish exposure. We also asked labs to recruit a sample of bilingual430
infants who shared at least one language – the community language being learned by431
monolinguals tested in the same lab. For labs in bilingual communities (e.g., Barcelona,432
Ottawa, Montréal, Singapore), labs were free to decide which community language to select
as the shared language. Within this constraint, most labs opted to test heterogeneous groups
of bilinguals, for example, English-Other bilinguals where English was the community435
language the other language might be French, Spanish, Mandarin, etc. Only one lab tested a
homogeneous group of bilinguals (in this case, all infants were learning English and437
Mandarin), although we had expected that more labs would test homogeneous samples,438
given both heterogeneous and homogeneous samples are used regularly in research with439
bilingual infants. Because only one homogeneous sample was tested, we were not able to440
conduct planned analyses examining whether the type of sample on our results. Infants that
were tested but that did not meet inclusion criteria into the group (for example because they
did not hear enough of their non-dominant language, or were not hearing the community443
language) were excluded from the main analyses, but retained for exploratory analyses where
Assessing bilingualism. Each lab was asked to use a detailed day-in-the-life446
parental interview questionnaire to quantify the percent of time that infants were exposed to
each language. This approach has been shown to predict bilingual children’s language448
outcomes better than a one-off parental estimate (DeAnda, Bosch, Poulin-Dubois, Zesiger, &
Friend, 2016). Moreover, recent findings based on day-long recording gathered using LENA450
technology show that caregivers can reliably estimate their bilingual child’s relative exposure
to each language (Orena et al., 2019). Labs were also asked to pay special attention to452
whether infants had exposure to North American English, and if so which caregiver(s) this453
input came from. As most of the labs contributing bilingual data had extensive expertise in
bilingual language background assessment, we encouraged each lab to use whatever version455
of measurement instrument was normally used in their lab (details of the assessment456
instruments are outlined below, including source references for most measures). Where457
possible, labs conducted the interview in the parents’ language of choice, and documented458
whether the parents’ preferred language was able to be used.459
While standardization of measurement tools is often desirable, we reasoned that460
different questions and approaches might be best for eliciting information from parents in461
different communities and from different cultures. Indeed, many labs reported that their own
instruments had undergone considerable refinement over the years as a function of their463
experience working with the families in their communities. However, in order to maximize464
the overall sample size and the diversity of bilingual groups tested, we encouraged465
participation from laboratories without extensive experience testing bilingual infants. Labs466
that did not have an established procedure were paired with more experienced labs working
with similar communities to refine a language assessment procedure. Twelve of the labs468
administered a structured interview-style questionnaire based on the one developed by Bosch
and Sebastián-Gallés (1997, 2001; for examples of the measure see the online supplementary
materials of Byers-Heinlein et al., 2019; DeAnda et al., 2016), and the remaining 5 labs471
administered other questionnaires. We describe each of these approaches in detail below.472
The Bosch and Sebastián-Gallés (1997, 2001) questionnaire is typically referred to in473
the literature as the Language Exposure Questionnaire (LEQ; e.g., Byers-Heinlein, Fennell,474
& Werker, 2013), or the Language Exposure Assessment Tool (LEAT; DeAnda et al., 2016).
Administration of these questionnaires takes the form of a parental interview, where a476
trained experimenter systematically asks at least one of the infant’s primary caregivers477
detailed questions about the infant’s language environment. The interviewer obtains an478
exposure estimate for each person who is in regular contact with the infant, as defined by a
minimum contact of once a week. For each of those people, the caregiver gives an estimate of
how many hours per day they speak to the infant in each language for each of the days of481
the week (e.g., weekdays and weekends may differ depending on work commitments).482
Further, the caregiver is asked if the language input from each regular-contact person was483
similar across the infant’s life history. If not, such as in the case of a caregiver returning to484
work after parental leave, or an extended stay in another country, an estimate is derived for
each different period of the infant’s lifespan. The interviewer also asks the caregiver about486
the language background of each person with regular contact with the infant (as defined487
above), asking the languages they speak and whether they are native speakers of those488
languages. The caregiver also gives an estimate of language exposure in the infant’s daycare,
if applicable. Finally, the caregiver gives a global estimate of their infant’s percent exposure
to the two languages, which includes input from those people in regular contact with the491
infant and other people with whom the infant has less regular contact (e.g., playgroups,492
friends of caregivers, etc.). Importantly, this global estimate does not include input from493
television or radio, as such sources have no known positive impact, and may even have a494
negative impact on monolingual and bilingual language development in infancy (see Hudon,
Fennell, & Hoftyzer, 2013). The estimate of an infant’s percent exposure to their languages
is derived from the average cumulative exposure based on the data from the primary497
individuals in the infant’s life. Some labs use the global estimate simply to confirm these498
percentages. Other labs average the primary and global exposure to take into account all499
language exposure, while still giving more weight to the primary individuals. Also, some labs
asked additional questions, for example about videoconferencing with relatives, whether501
caregivers mix their languages when speaking to the infant, or caregivers’ cultural502
background. Finally, while the original form was pen-and-paper, there have been adaptations
which include using a form-fillable Excel sheet (DeAnda et al., 2016).504
For the other language exposure measures used by 5 of the labs, we will simply505
highlight the differences from the LEQ/LEAT measure described above, as there is much506
overlap between all the instruments used to measure infants’ exposure to their languages.507
Two labs used custom assessment measures designed within each lab. The major difference508
from the LEQ for the first of these custom measures is that parents provide percentage509
exposure estimates for each language from primary individuals in the infant’s life, rather510
than exposure estimates based on hours per day in each language. The other custom511
measures, unlike the LEQ, specifies estimates of language exposure in settings where more512
than one speaker is present by weighting each speaker’s language contribution. A further two
labs used other child language exposure measures present in the literature: one used the514
Multilingual Infant Language Questionnaire (MILQ; Liu & Kager, 2017b) and the other used
an assessment measure designed by Cattani et al. (2014). For the MILQ, one major516
difference is that parents complete the assessment directly using an Excel sheet with clear517
instructions. The other major difference is that the MILQ is much more detailed than the518
LEQ/LEAT: breaking down language exposure to very specific activities (e.g., car time, book
reading, meal time); asking more detail about the people in regular contact with the infant520
(e.g., accented speech, level of talkativeness); and obtaining estimates of media exposure521
(e.g., TV, music). The measure from Cattani et al. (2014) focuses on parental exposure and
uses Likert scales to determine exposure from each parent. The ratings are converted to523
percentages and maternal exposure is weighted more in the final calculation based on data524
showing that mothers are more verbal than fathers. Finally, one lab did not use a detailed525
measure, but rather simply asked parents to give an estimate of the percentage exposure to
each of the languages their infant was hearing.527
We asked labs that collected data from monolingual infants (a subset of the data528
included in the ManyBabies 1 monolingual study) to check participants’ monolingual status
by asking parents a single question: estimate the percent of time that their infant was530
exposed to their native language. If that estimate exceeded 90% exposure to a single531
language, the infant was considered monolingual.532
Demographics. Each lab administered a questionnaire that gathered basic533
demographic data about infants, including age, health history, gestation, etc. Infants’534
socioeconomic status (SES) was measured via parental report of years of maternal education.
To standardize across different education systems where formal schooling may begin at536
different ages, we counted the number of years of education after kindergarten. For example,
in the United States, mothers who had completed high school would be considered to have538
12 years of education.539
Final sample. Our final sample of bilinguals who met our infant-level inclusion540
criteria included 333 infants tested in 17 labs; 148 were 6–9 months, and 185 were 12–15541
months (full account of exclusions is detailed in the results section). These 17 labs also542
collected data from monolingual infants (N = 385 who met infant-level inclusion criteria), of
whom 182 were 6–9 months, and 203 were 12–15 months. While all analyses required that544
data meet the infant-level inclusion criteria, some analyses further required that the data met
the lab-level inclusion criteria (lab-level inclusion criteria are discussed in the Results section
where they were implemented for specific analyses). Data from monolingual infants in these
age ranges were available from 59 additional labs (n = 583 6-9 month-olds; n = 468 12-15548
month-olds) who did not contribute bilingual data. Bilingual infants and their lab-matched
monolingual samples tested by each lab are detailed in Table 1. For further description of550
our participants, please refer to the Appendix, where we list gender distributions across551
subsamples (Table A1) and the language pairs being learned by bilingual infants (Table A2).
Additional information about the larger ManyBabies sample from which the monolingual553
data were sampled can be found in the ManyBabies 1 monolingual study (ManyBabies554
Consortium, in press) and the associated Open Science Framework project at
Table 1
Number of monolingual and bilingual infants in each age group by lab which met infant-level
inclusion criteria. Note that because of lab-level inclusion criteria, cells with n < 10 were
excluded from the meta-analytic analyses, but were included in the mixed-effects regression
analyses. Labs that only tested monolingual infants are not listed.
lab method 6-9 mo
6-9 mo
12-15 mo
12-15 mo
babylabbrookes singlescreen 17 15 17 16
babylabkingswood hpp 9 15 15 15
babylabparisdescartes1 hpp 10 0 1 16
babylabprinceton hpp 15 1 0 0
bllumanitoba hpp 7 26 8 16
cdcceu eyetracking 0 0 14 13
infantcogubc eyetracking 10 11 0 0
infantstudiesubc hpp 15 20 0 0
irlconcordia eyetracking 16 17 18 18
isplabmcgill hpp 0 0 16 11
langlabucla hpp 0 0 9 3
ldlottawa singlescreen 7 17 18 11
lllliv eyetracking 7 19 6 15
lscppsl eyetracking 0 0 16 14
nusinfantlanguagecentre eyetracking 26 10 12 10
weltentdeckerzurich eyetracking 0 0 28 30
wsigoettingen singlescreen 9 31 7 15
Visual stimuli. Labs using a central fixation or eye-tracking method presented557
infants with a brightly-coloured checkerboard as the main visual stimulus. A video of a558
laughing baby was used as an attention-getter between trials to reorient infants to the screen.
Labs using the headturn preference procedure used the typical visual stimulus employed in560
their labs, which was sometimes light bulbs (consistent with the original development of the
procedure in the 1980s) or sometimes colourful stimuli presented on LCD screens. All visual
stimuli are available via the ManyBabies 1 monolingual Open Science Framework site at563
Auditory stimuli. Auditory stimuli consisted of semi-naturalistic recordings of565
mothers interacting with their infants (ranging in age from 122–250 days) in a laboratory566
setting. Mothers were asked to talk about a set of objects with their infant, and also567
separately with an experimenter. A set of 8 IDS and 8 ADS auditory stimuli of 20 s each568
were created from these recordings. Details regarding the recording and selection process,569
acoustic details and ratings from naive adult listeners can be found in the ManyBabies 1570
monolingual study (ManyBabies Consortium, in press) and the associated Open Science571
Framework project at
Basic Procedure. Each lab used one of three common infant study procedures,574
according to their own expertise and the experimental setups available in the lab: central575
fixation (3 labs), eye-tracking (7 labs), or headturn preference procedure (7 labs). The576
testing procedure was identical to that used in the ManyBabies 1 monolingual project577
(ManyBabies Consortium, in press, deviations from the protocol are also described there),578
and only key aspects will be briefly summarized here.579
Infants sat on their parents’ laps or in a high chair, and parents listened to masking580
music over headphones throughout the study. Infants saw 2 training trials that presented an
unrelated auditory stimulus (piano music), followed by 16 test trials that presented either582
IDS or ADS speech. Trials were presented in one of four pseudo-random orders that583
counterbalanced the order of presentation of the two stimulus types. Note that within each584
order, specific IDS and ADS clips were presented adjacently in yoked pairs to facilitate585
analyses. On each trial, the auditory stimulus played until the infant looked away for 2586
consecutive seconds (for labs that implemented an infant-controlled procedure) or until the587
entire stimulus played, up to 19 seconds (for labs that implemented a fixed trial-length588
procedure). The implementation of the procedure depended on the software that was589
available in each lab. Trials with less than 2 seconds of looking were excluded from analyses.
Attention-grabbing stimuli were played centrally between trials to reorient infants to the task.
The main differences between the setups were the type and position of visual stimuli592
presented, and the onset of the auditory stimuli. For central fixation and eye-tracking593
procedures, infants saw a checkerboard on a central monitor, whose presentation coincided594
with the onset of the auditory stimuli on each trial. For the headturn preference procedure,595
the visual stimulus (either flashing light bulbs or a colourful stimulus) played silently on a596
monitor/bulb in the centre of the room and on one of two side monitors/bulbs, and the597
auditory stimulus began playing when the infant turned their head towards the side stimulus.
The dependent variable was infant looking time during each trial. For eye-tracking599
setups, looking time was measured automatically via corneal reflection. For central fixation
and headturn preference procedure setups, looking time was measured by trained human601
coders who were blind to trial type, according to the lab’s standard procedures.602
Parents completed questionnaires about participants’ demographic and language603
background either prior to or after the main experiment.604
Analysis overview606
Data exclusion. Labs were asked to submit all data collected as part of the607
bilingual study to the analysis team, and this section focuses on exclusions for infants608
collected as part of the bilingual sample. The initial dataset contained 501 bilingual infants,
of which 333 met each of the following inclusion criteria. These criteria are detailed below.610
We note that exclusions were applied sequentially (i.e., percentages reflect exclusions among
remaining sample after previous criteria were applied).612
Age. We included infants in two age groups: 6-9 and 12-15 month-olds. There were 59
(11.78%) bilingual infants who were tested in the paradigm, but who fell outside our614
target ages.615
Bilingualism. We excluded infants from the bilingual sample whose language616
background did not meet our pre-defined criteria for bilingualism (see above for617
detials). There were 74 (16.74%) infants whose exposure did not meet this criterion.618
We also excluded an additional 7 (1.90%) infants who met this criterion, but who were
not learning the community language as one of their languages.620
Full-term. We defined full term as gestation times greater than or equal to 37 weeks.621
There were 4 (1.11%) bilingual infants who were tested but did not meet this criterion.
No diagnosed developmental disorders. We excluded infants whose parents reported623
developmental disorders (e.g., chromosomal abnormalities, etc.) or were diagnosed with
hearing impairments. There were 2 (0.56%) infants who were tested but did not meet
this criterion. Due to concerns about the accuracy of parent reports, we did not plan626
exclusions based on self-reported ear infections unless parents reported627
medically-confirmed hearing loss.628
Session-level errors. Participants were also excluded on the basis of session-level errors,
including 2 infants for equipment error, 1 infants for experimenter error and 3 infants630
for outside interference.631
Adequate trials for analysis. We excluded any infant who did not have at least one632
IDS-ADS trial pair available for analysis: 5 (1.45%) infants were tested but did not633
meet these criteria. For infants with at least one good trial pair, we additionally634
excluded any trial with less than 2 s of looking (n = 876; 16.92%), which was set as a
trial-level minimum so that infants had heard enough of the stimulus to discriminate636
IDS from ADS. As infants did not have to complete the entire experiment to be637
included, this meant that different infants contributed different numbers of trials.638
Data analysis framework. All planned analyses were pre-registered at639; data and code are available at640 Our primary dependent variable of641
interest was looking time (LT), which was defined as the time spent fixating on the visual642
stimulus during test trials. Given evidence that looking times are non-normally distributed,
we log-transformed all looking times prior to statistical analysis (Csibra, Hernik, Mascaro,644
Tatone, & Lengyel, 2016). We refer to this transformed variable as “log LT”. We645
pre-registered a set of analyses to examine whether monolinguals, heterogeneous samples of
bilinguals, and homogeneous samples of bilinguals showed different levels of variability.647
Unexpectedly, only 1 lab (Table 1) tested a homogenous sample of bilinguals, thus we648
deviated from our original plan and did not analyze data as a function of whether our649
bilingual groups were homogenous versus heterogeneous. For the main analyses, we adopted
two complementary data analytic frameworks parallel to the ManyBabies 1 monolingual651
project (ManyBabies Consortium, in press): meta-analysis and mixed-effects regression.652
Under the meta-analytic framework, data from each sample of infants (e.g., 6 to 9653
month-old bilinguals from Lab 1) was characterized by a) its effect size (here Cohen’s
), and
b) its standard deviation. Effect size analyses addressed questions about infants’ overall655
preference for IDS, while group-based standard deviation analyses addressed questions about
whether some groups of infants show higher variability in their preference than others. Note
that meta-analyses of intra-group variability are relatively rare (Nakagawa et al., 2015;658
Senior, Gosby, Lu, Simpson, & Raubenheimer, 2016). For the meta-analysis only, we659
implemented a lab-level inclusion criterion, such that each effect size was computed only if660
the lab had contributed at least 10 infants in that particular language group and age. For661
example, if lab A had contributed 7 bilingual infants between 6- to 9-months and 15662
monolingual infants between 6- to 9-months, we only computed effect size for the663
monolingual group, but not for the bilingual group. This criterion ensured that each effect664
size was computed based on a reasonable sample size (i.e., a minimum of 10 infants) and also
was consistent with the lab-level inclusion criteria in the ManyBabies 1 monolingual study.666
An advantage of the meta-analytic approach is that it is easy to visualize lab-to-lab667
differences. Further, the meta-analytic framework most closely mirrors the current approach
for studying monolingual-bilingual differences, which typically compares groups of669
monolingual and bilingual infants tested within the same lab. We used this approach670
specifically to test the overall effect of bilingualism and its possible interactions with age on
the magnitude of infants’ preference for IDS over ADS. We also compared standard672
deviations for the bilingual group and monolingual group in a meta-analytic approach. This
analysis closely followed Nakagawa et al. (2015).674
Under the mixed-effects regression model, trial-by-trial data from each infant were675
submitted for analysis. Further, independent variables of interest could be specified on an676
infant-by-infant basis. This approach had the advantage of potentially increasing statistical
power, as data are analyzed at a more fine-grained level of detail. As with the meta-analytic
approach, this analysis tested the effects of bilingualism and their potential interactions with
age. We also investigated whether links between bilingualism and IDS preference were680
mediated by socio-economic status. Additionally, this approach allowed us to assess how the
amount of exposure to NAE-IDS, measured as a continuous percentage, affected infants’682
listening preferences. Note that unlike for the meta-analysis, we did not apply a lab-level683
inclusion criterion in order to maximize our sample size. Thus, data from all infants who met
the infant-level criteria were included in this analysis, resulting in slightly different sample685
sizes under the meta-analytic and mixed-effects approaches.686
Under both frameworks, we used a dual analysis strategy to investigate how infants’687
IDS preference is related to bilingualism. First, we examined the matched subset of data688
from labs that contributed a monolingual and bilingual sample at a particular age. Second,689
we examined the complete set of data including data from labs that contributed both690
monolinguals and bilinguals, as well as additional data from labs that only collected691
monoinguals at the ages of interest as part of the larger ManyBabies 1 project.692
Confirmatory analysis section693
Meta-analytic approach.. This approach focused on the analysis of group-level694
datasets. We defined a dataset as a group of at least 10 infants tested in the same lab, of the
same age (either 6-9 or 12-15 months), and with the same language background696
(monolingual or bilingual). For analyses of within-group variability, we compared bilingual697
infants to monolingual infants.698
To estimate an effect size for each dataset, we first computed individual infants’699
preference for IDS over ADS by 1) subtracting looking time to the ADS stimulus from700
looking time to the IDS stimulus within each yoked trial pair, and 2) computing a mean701
difference score for each infant. Pairs that had a trial with missing data were excluded702
(42.93% pairs in matched dataset, 40.34% pairs in full dataset), which constituted a total of
30.77% of trials in matched dataset, 31.02% of trials in full dataset. Note that we expected704
many infants to have missing data particularly on later test trials, given the length of the705
study (16 test trials). Then, for each dataset (i.e., combination of lab, infant age group, and
whether the group of participants was bilingual or monolingual), we calculated the mean of
these difference scores (Md) and its associated standard deviation across participants (sd).708
Finally, we used the derived Mdand sd to compute a within-subject Cohen’s dusing the709
formula dz=Md/sd.710
In the following meta-analyses, random effects meta-analysis models with a restricted711
maximum-likelihood estimator (REML) were fit with the metafor package (Viechtbauer,712
2010). To account for the dependence between monolingual and bilingual datasets stemming
from the same lab, we added laboratory as a random factor. As part of our pre-registered714
analyses, we planned to include method as a moderator in this analysis if it was found to be
a statistically significant moderator in the larger ManyBabies 1 monolingual project - which
it was (ManyBabies Consortium, in press). However, because only 17 labs contributed717
bilingual data, we deviated from this plan because of the small number of labs per method718
(e.g., only three labs used a single-screen method).719
Effect size-based meta-analysis..
Our first set of meta-analyses focused on effect
sizes (
): how our variables of interest contributed to effect size comparing looking time to
IDS versus ADS trials. As a reminder, we ran the analyses in two ways: (i) the analysis was
only restricted to the labs that contributed matched data (matched dataset), and (ii) the723
analysis included all available data labs that tested only monolinguals or only bilinguals at724
the ages of interest (full dataset).725
We initially fit the following model to examine contributions of age and bilingualism to
infants’ IDS preference, as well as potential interactions between these variables:727
dz1 + bilingual +age +bilingual * age
Bilingualism was dummy coded (0 = monolingual, 1 = bilingual), and age was coded728
as the average age for each lab’s contributed sample for each language group (centered for729
ease of interpretation).730
In the matched dataset, we did not find any statistically significant effects of age (
0.17, CI = [-1.01 - 1.36], z = 0.29,
775), bilingualism (
= -0.17, CI = [-0.44 - 0.10], z
= -1.22, p=.224), or interactions between age and bilingualism (dz= -0.19, CI = [-1.84 -733
1.46], z = -0.22, p=.822).734
Similarly, in the full dataset, we did not find any significant main effects of age, (
0.01, CI = [-0.65 - 0.67], z = 0.02,
982), bilingualism (
= -0.10, CI = [-0.29 - 0.09], z
= -1.04,
299), nor a significant interaction between age and bilingualism (
= 0.01, CI
= [-0.93 - 0.95], z = 0.02, p=.981).738
As bilingualism is the key moderator of research interest in the current paper, here we
report the effect sizes of monolingual and bilingual infants separately. In the matched740
dataset, the effect size for monolinguals was dz= 0.42 (CI = [0.21 - 0.63], z = 3.94,741
p=< .001), while for bilinguals the effect was dz= 0.24 (CI = [0.06 - 0.42], z = 2.64,742
p=.008). In the full dataset, the effect size for monolinguals was dz= 0.36 (CI = [0.28 -743
0.44], z = 9.20,
< .
001), while for bilinguals the effect was
= 0.26 (CI = [0.09 - 0.43], z
= 2.97, p=.003). In sum, numerically monolinguals showed a stronger preference for IDS745
than bilinguals, but this tendency was not statistically significant in the effect size-based746
meta-analyses. A forest plot for this meta-analysis is shown in Figure 1.747
Within-group variability meta-analysis. Our second set of pre-registered748
meta-analyses examined whether the variability in infants’ preference for IDS within a749
sample (within-study variability) was related to language background (monolingual750
vs. bilingual). Note that this question of within-sample heterogeneity is different than751
questions of between-sample heterogeneity that can also be addressed in meta-analysis (see752
Higgins & Thompson, 2002; Higgins, Thompson, Deeks, & Altman, 2003 for approaches to753
6−9 mo
12−15 mo
−1 0 1 2
Meta−analytic estimate
Effect Size
Language group Bilingual infants Monolingual infants
Figure 1 . Forest plot for the matched dataset. Standardized effect sizes are shown for
each lab, with error bars showing 95% confidence intervals. Each lab reported two effect
sizes: one for monolingual group and the other one for bilingual group. Points are scaled
by inverse variance and colored by language groups (blue circle denotes bilinguals and red
triangle denotes monolinguals). For the age group panels, the point and its associated interval
represent the meta-analytic estimate and the 95% confidence interval in each language group.
The points in bottom panel show the global meta-analytic estimate.
between-group variability in meta-analysis). Specifically, the within-group variability754
meta-analysis approach provides additional insights of how two groups differ in terms of their
variances, not merely their mean effect sizes. This approach is useful when the language756
backgrounds of the infants influence not only the magnitude of infants’ IDS preference, but757
also the variability of infants’ IDS preference. In the following, the standard deviations758
measure looking time variability of infants’ preference for IDS over ADS in each language759
group (either monolingual or bilingual). Again, we report
, an effect size that measures the
magnitude of infants’ preference for IDS over ADS.761
According to Nakagawa et al. (2015), there are two approaches to run within-group762
variability meta-analysis: one approach uses lnCV R, the natural logarithm of the ratio763
between the coefficients of variation, to compare the variability of two groups; a second764
approach enters lnSD (the natural logarithm of standard deviations) and ln ¯
X(the log765
mean) into a mixed-effect model. The mixed-effect model can be used when the assumption
that the standard deviation is proportional to the mean, necessary for the
lnCV R
is not met. We found that the standard deviations and effect sizes were not significantly768
correlated. Therefore, we chose the second, mixed-effect model approach. In the following769
meta-regression model, the natural logarithm of the standard deviations (lnSD) from each770
language group is the dependent variable. This dependent variable (group variance) is the771
log-transformed standard deviation of infants’ preference for IDS over ADS that correspond
to infants’ language group (either monolingual/bilingual).773
lnSD 1 + bilingual +ln(d0
is the absolute value of
because we needed to ensure that values entered into the
logarithm were positive, bilingual is the binary dummy variable that indicates bilingualism -
whether the language group is monolingual or bilingual. Further, bilingualism was entered as
a random slope in the model.777
In the matched dataset, we did not find statistically significant evidence for778
bilingualism as a moderator of the differences in standard deviations across language groups,
= -0.08,
235). Similarly, we also did not find statistical significance for bilingualism
in the full dataset, (
= 0.03,
660). In short, we did not find support for the hypothesis
that bilingual infants would show larger within-group variability than monolingual infants.782
Mixed-effects approach.
Mixed-effects regression allows variables of interest to be
specified on a trial-by-trial and infant-by-infant basis. We had anticipated that we would be
able to include additional data from labs that aimed to test homogeneous samples (i.e.,785
because we could include infants from these labs who were not learning this homogeneous786
language pair), but in practice this did not apply as only one lab contributed a homogeneous
data set, and that lab did not test additional infants. We were also able to include data from
all valid trials, rather than excluding data from yoked pairs with a missing data point. As789
under the meta-analytic approach, we ran the models twice, once including only data from790
labs that contributed matched samples of monolinguals and bilinguals, and once including all
available data from 6-9 and 12-15 month-olds.792
The mixed-effects model was specified as follows:793
DV IV1+IV2+... + (...|subject)+(...|item)+(...|lab)
The goal of this framework was to examine effects of the independent variables (IV) on
the dependent variable (DV), while controlling for variation in both the DV (“random795
intercepts”) and the relationship of the IV to the DV (“random slopes”) based on relevant796
grouping units (subjects, items, and labs). Following recent recommendations (Barr, Levy,797
Scheepers, & Tily, 2013), we planned to initially fit a maximal random effects structure, such
that all random effects appropriate for our design were included in the model. However, we
also recognized that such a large random effects structure might be overly complex given our
data, and would be unlikely to converge. After reviewer feedback during Stage 1 of the801
Registered Report review process, we pre-registered a plan to use a “Parsimonious mixed802
models” approach for pruning the random effects (Bates et al., 2015a; Matuschek, Kliegl,803
Vasishth, Baayen, & Bates, 2017). However, we found that it was computationally difficult804
to first fit complex models (i.e., our models had multiple interactions and cross-levels805
grouping) under the maximal random effects structure and then prune the models using a806
parsimonious mixed models approach. Further, we note that this was not the approach used
in MB1, which would make direct comparison between MB1 and the current study difficult.
As such, following MB1, we fitted and pruned1the following models using the maximal809
random effects structure only (Barr et al., 2013). We fit all models using the lme4 package810
(Bates et al., 2015b) and computed pvalues using the lmerTest package (Kuznetsova,811
Brockhoff, & Christensen, 2016). Below is a description of our model variables:812
log_lt: Dependent variable. Log-transformed looking time in seconds.813
trial_type: A dummy coded variable with two levels, with ADS trials as the baseline,814
such that positive effects of trial type indicate longer looking to IDS.815
bilingual: A dummy coded variable with two levels, with monolingual as the baseline,816
such that positive effects of bilingualism reflect longer looking by bilinguals.817
language: A dummy coded variable for whether infants were learning North American
English as a native language (i.e., >= 90% exposure to NAE for monolinguals, or >=
25% exposure to NAE for bilinguals).820
exp_nae: A continuous variable for the percent of time infants heard North-American
method: A dummy-coded variable to control for effects of different experimental823
setups, with single-screen central fixation as the reference level.824
age_days: Centered for interpretability of main effects.825
trial_number: The number of the trial pair, recoded such that the first trial pair is 0.826
1Results reported in this paper were pruned by fitting mixed-effect models with ‘lme4‘ version 1.1-21.
ses: The number of years of maternal education, centered for ease of interpretation.827
Note that in this analysis plan, we have used a concise format for model specification,828
which is the form used in R. As such, lower-order effects subsumed by interactions are829
modeled even though they are not explicitly written. For example, the interaction trial_type
* trial_num also assumes a global intercept, a main effect of trial type, and a main effect of
trial number.832
Homogeneity of variance.
We pre-registered a Levene’s test to examine whether
monolinguals and bilinguals showed different amounts of variance in their IDS preference.834
However, we ultimately omitted this test because the null results of our within-group835
variability meta-analysis did not support a difference between monolingual and bilingual836
infants in the variance of their IDS preference.837
Effects of bilingualism on IDS preference.
We planned a mixed-effects model
which was based on the structure of the final model fit for the ManyBabies monolingual839
project, including bilingualism as an additional moderator. Note that because data collection
for both projects was simultaneous, we did not know prior to registration what the final841
model structure for the monolingual-only sample would be (it was expected that pruning of
this model would be necessary in the case of non-convergence). The original model proposed
for the monolingual-only sample was designed to include simple effects of trial type, method,
language (infants exposed vs. not exposed to NAE-IDS), age, and trial number, capturing845
the basic effects of each parameter on looking time (e.g., longer looking times for IDS,846
shorter looking times on later trials). Additionally, the model included two-way interactions
of trial type with method and with trial number, a two-way interaction of age with trial848
number, as well as two- and three-way interactions between trial type, age, and language849
(see ManyBabies Consortium, in press, for full justification). This model was specified to850
minimize higher-order interactions while preserving theoretically-important interactions.851
Note that to reduce model complexity, both developmental effects and trial effects are852
treated linearly. The planned model for ManyBabies monolingual was:853
log lt trial type method +trial type trial num +age trial num+
trial type age language+
(trial type trial num |subid)+
(trial type age |lab)+
(method +age language |item)
It was expected that pruning would be necessary in the case of non-convergence.854
Our analysis plan specified that we would add bilingualism to the fixed effects of the855
final pruned model that fitted to the monolingual sample. For higher-order interactions in856
the model, we ensured that we had at least 20 infants per group. For example, for a857
three-way interaction between bilingualism, language and age, we included at least 20 infants
per group: at least 20 infants in the group of 6-9 month-old bilinguals who were not exposed
to NAE. We applied the same rules to all other groups.860
In our preregistration, we were uncertain as to whether our sample size would support
a model with a four-way-interaction of trial type, age, bilingual status, and language. Given
our final sample size, we elected to fit our main model without including the four-way863
interaction effect
. In our main model, we included two fixed three-way interactions: (i) the
interaction between bilingualism, age and trial type, and (ii) the interaction between865
language, age and trial type, as well as other subsumed lower-order interactions.866
Regardless of our fixed effect structure, the model included the random slope of867
We did not enter the above-mentioned four-way interaction into our main model, but note that in the
more complex model, the four-way interaction was not statistically significant in the matched dataset (
0.00, SE = 0.02,p= 0.85) or the full dataset (β= 0.01, S E = 0.01,p= 0.63).
bilingualism on lab and item, as well as appropriate interactions with other random factors.
Our initial unpruned model was:869
log lt trial type method +trial type trial num +age trial num+
trial type age language+
trial type age bilingual+
(trial type trial num |subid)+
(trial type age bilingual |lab)+
(method +age language +age bilingual |item)
After pruning random effects for non-convergence and singularity, the final models for
the matched dataset and full dataset were different. The following was the final model of the
matched dataset:872
log lt trial type method +trial type trial num +age trial num+
trial type age language+
trial type age bilingual+
(1 |subid)+
(bilingual |lab)+
(1 |item)
In contrast, the final model of the full dataset was:873
log lt trial type method +trial type trial num +age trial num+
trial type age language+
trial type age bilingual+
(1 |subid)+
(1 |lab)+
(1 |item)
Overall, the mixed-level analyses in both matched and full datasets yielded similar874
results (Table 2 and 3). More coefficients were statistically significant in the full dataset,875
likely due to the larger sample size. Thus, in the following, we focus on the results of the876
mixed-level model for the full dataset. We found that infants showed a preference for IDS, as
indicated by a positive coefficient on the IDS predictor (reflecting greater looking times to878
IDS stimuli). We did not find any effects of the bilingualism on IDS preference nor any879
interaction effects between bilingualism and other moderators. This finding is consistent880
with the results of our meta-analysis above.881
Surprisingly, the fitted model did not show an interaction between infants’ IDS882
preference and the method used in the lab, a result that is different from the results in the883
MB1 project. However, this finding is likely due to smaller sample sizes in the current paper,
as we restricted the analysis to participants at particular ages. Apart from this, our findings
were largely consistent with the MB1 study. There was a significant and positive two-way886
interaction between IDS and NAE, suggesting greater IDS preferences for children in NAE887
contexts. The interaction between IDS and age was also significant and positive, suggesting
that older children showed a stronger IDS preference. Finally, we found a marginally889
significant three-way interaction between IDS, age, and NAE, suggesting that older children
in NAE contexts tended to show stronger IDS preference than those in the non-NAE891
Table 2
Linear Mixed Model 1 testing bilingualism effect on IDS in a
matched dataset.
Estimate SE t p
Intercept 1.930 0.074 26.000 0.000
IDS 0.093 0.047 2.000 0.050
HPP 0.103 0.092 1.110 0.283
Single Screen 0.113 0.103 1.090 0.288
Age -0.027 0.008 -3.410 0.001
Trial # -0.036 0.003 -13.900 0.000
NAE -0.059 0.075 -0.792 0.435
Bilingual 0.000 0.034 0.008 0.994
IDS * HPP 0.016 0.029 0.566 0.571
IDS * Single Screen 0.004 0.031 0.124 0.901
Age * Trial # 0.001 0.000 2.270 0.023
IDS * Trial # 0.001 0.004 0.175 0.862
IDS * Age 0.013 0.006 2.180 0.029
IDS * NAE 0.051 0.026 1.950 0.052
Age * NAE 0.007 0.010 0.646 0.519
IDS * Bilingual -0.012 0.024 -0.522 0.602
Age * Bilingual -0.006 0.009 -0.671 0.503
IDS * Age * NAE 0.016 0.008 1.860 0.063
IDS * Age * Bilingual -0.009 0.008 -1.210 0.227
Table 3
Linear Mixed Model 1 testing bilingualism effect on IDS in a
full dataset.
Estimate SE t p
Intercept 1.890 0.047 40.400 0.000
IDS 0.106 0.038 2.770 0.009
HPP 0.190 0.058 3.310 0.002
Single Screen 0.243 0.054 4.510 0.000
Age -0.029 0.005 -5.680 0.000
Trial # -0.037 0.002 -21.200 0.000
NAE 0.003 0.048 0.063 0.950
Bilingual -0.006 0.025 -0.234 0.815
IDS * HPP 0.029 0.018 1.620 0.106
IDS * Single Screen -0.020 0.019 -1.060 0.291
Age * Trial # 0.001 0.000 3.910 0.000
IDS * Trial # -0.002 0.002 -0.961 0.337
IDS * Age 0.013 0.003 3.800 0.000
IDS * NAE 0.038 0.016 2.420 0.015
Age * NAE 0.002 0.007 0.244 0.807
IDS * Bilingual 0.003 0.019 0.142 0.887
Age * Bilingual -0.003 0.008 -0.369 0.712
IDS * Age * NAE 0.009 0.005 1.960 0.051
IDS * Age * Bilingual -0.007 0.006 -1.110 0.265
Dose effects of exposure to NAE-IDS in bilingual infants.
In this analysis,
we tested whether we could observe a dose-response relationship between infants’ exposure894
to NAE-IDS (measured continuously) and their preference for IDS over ADS.895
We decided to conduct this analysis only including data from bilinguals. Our reasoning
was that bilingualism status and exposure to NAE-IDS are confounded, as monolinguals’897
exposure to NAE will be either near 0% or 100%, while bilinguals’ NAE experience can be898
either 0%, or 25-75%. Because the monolingual sample is larger and their NAE exposures899
are more extreme, their effects would dominate that of the bilinguals in a merged analysis.900
Therefore, we reasoned that if there is a dose effect, it should be observable in the bilingual
sample alone. Finally, although excluding monolingual infants reduced power overall, we902
decided that given the relatively large sample of bilingual infants, this disadvantage would be
offset by the ease of interpretation afforded by restricting the analysis to bilinguals.904
Once again, we based this model on the final pruned monolingual model, substituting
the binary measure of exposure to NAE-IDS (language) with the continuous measure of906
exposure(exp_nae), and including a random slope for exp_nae by item (which was907
ultimately pruned from the model). After pruning, our model was specified as follows:908
log lt trial type method +trial type trial num +age trial num+
trial type age exp nae+
(1 |subid)+
(trial type |lab)+
(1 |item)
Table 4 contains the details of the results in this model. The main effect of infants’909
exposure to NAE (exp_nae) was not significant (
= -0.00067,
= 0
= 0
57). This
indicates that bilingual infants who were exposed to more NAE did not pay more attention
Table 4
Linear Mixed Model testing the effects of exposure to NAE-IDS
in bilingual infants.
Estimate SE t p
Intercept 1.910 0.074 25.900 0.000
IDS -0.009 0.062 -0.138 0.891
HPP 0.088 0.091 0.963 0.354
Single Screen 0.168 0.111 1.510 0.160
Age -0.024 0.010 -2.270 0.024
Trial # -0.036 0.004 -10.100 0.000
EXP_NAE -0.001 0.001 -0.565 0.575
IDS * HPP 0.054 0.053 1.020 0.331
IDS * Single Screen 0.028 0.060 0.465 0.654
Age * Trial # 0.000 0.001 0.300 0.764
IDS * Trial # 0.006 0.005 1.150 0.251
IDS * Age 0.006 0.008 0.781 0.435
IDS * EXP_NAE 0.002 0.001 2.860 0.011
Age * EXP_NAE 0.000 0.000 -0.200 0.842
IDS * Age * EXP_NAE 0.000 0.000 0.891 0.373
to the NAE speech stimuli than those who were exposed to less NAE. However, the912
interaction between trial type and exp_nae was significant (β= 0.0023, SE = 0.00081,913
p= 0.011). That is, bilingual infants who were exposed to more NAE showed stronger IDS914
preferences, confirming a dose-response relationship between infants’ exposure to NAE and915
their preference for IDS over ADS (Figure 2) even among bilinguals who are learning NAE916
as one of their native languages.917
0 20 40 60
NAE exposure (%)
IDS preference (s)
Model fit all data partial data (25%−75% nae exposure)
Figure 2 . Linear trend between infants’ IDS preference and their percentage of time exposed
to NAE. Blue line indicates a regression model between infants’ IDS preference and their
NAE exposure (starting from zero). Red line indicates another regression model of the same
relationship with a focus of NAE exposure between 25 to 75%. Finally, we note that the
y-axis was truncated to highlight the trend such that some individual points are not plotted.
Socio-economic status as a moderator of monolingual-bilingual918
Because socio-economic status can vary systematically between monolinguals
and bilinguals in the same community, we were interested in whether relationships between920
bilingualism and IDS preference would hold when controlling for socio-economic status. It is
possible that an observed effect of bilingualism on IDS preference could disappear once SES
was controlled. Alternatively, it is possible that the effect of bilingualism on IDS preference
could only be apparent once SES was controlled. Thus, this analysis was important924
regardless of an observed relationship between IDS preference and bilingualism in the925
previous model.926
Our approach was to add SES as a moderator of our final model for bilinguals. We927
expected that any effects of socio-economic status could interact with age, thus this model928
included interactions of trial type, age, and socio-economic status as a fixed effect, as well as
the corresponding random slope by item. Based on the potential model detailed above for930
the bilinguals, our expected ses-mediated model was:931
log lt trial type method +trial type trial num +age trial num+
trial type age language+
trial type age bilingual+
trial type age ses+
(trial type trial num |subid)+
(trial type age bilingual |lab)+
(method +age language +age bilingual +age ses |item)
After pruning for non-convergence, our final model specifications are listed below. For
the matched dataset, the final model was:933
log lt trial type method +trial type trial num +age trial num+
trial type age language+
trial type age bilingual+
trial type age ses+
(1 |subid)+
(bilingual |lab)
By contrast, the final model of the full dataset was:934
log lt trial type method +trial type trial num +age trial num+
trial type age language+
trial type age bilingual+
trial type age ses+
(1 |subid)+
(1 |lab)+
(1 |item)
In general, across the matched and full datasets (Table 5 and 6), SES did not have a935
significant effect on infants’ looking time nor did it affect infants’ preference for IDS.936
However, for the matched dataset only, we found a statistically significant three-way937
interaction between IDS, age, and SES. Specifically, infants from 6- to 9-month-olds showed
stronger IDS preference when they were from a higher SES families, but older infants from939
12- to 15-month-olds showed similar IDS preference across families with different SES levels.
However, this interaction was not observed in the full dataset, raising the possibility that it
is a spurious, and arose only in the matched dataset because it is substantially smaller than
the full data set.943
Table 5
Linear Mixed Model 3 examining socio-economic status as a
moderator of monolingual-bilingual differences SES in the
matached dataset.
Estimate SE t p
Intercept 1.960 0.108 18.200 0.000
IDS 0.075 0.080 0.936 0.349
HPP 0.120 0.089 1.340 0.199
Single Screen 0.094 0.100 0.939 0.359
Age -0.025 0.029 -0.881 0.378
Trial # -0.033 0.002 -17.200 0.000
NAE -0.089 0.072 -1.240 0.225
Bilingual 0.022 0.028 0.795 0.427
SES -0.003 0.005 -0.513 0.608
IDS * HPP 0.019 0.030 0.633 0.527
IDS * Single Screen 0.006 0.032 0.201 0.841
Age * Trial # 0.001 0.000 2.330 0.020
IDS * Trial # -0.005 0.003 -1.740 0.081
IDS * Age 0.070 0.025 2.770 0.006
IDS * NAE 0.054 0.028 1.960 0.050
Age * NAE 0.012 0.010 1.130 0.260
IDS * Bilingual -0.018 0.025 -0.734 0.463
Age * Bilingual -0.010 0.009 -1.160 0.246
IDS * SES 0.003 0.005 0.770 0.441
Age * SES 0.000 0.002 -0.147 0.883
IDS * Age * NAE 0.016 0.009 1.810 0.071
IDS * Age * Bilingual -0.005 0.008 -0.606 0.545
IDS * Age * SES -0.004 0.002 -2.330 0.020
Table 6
Linear Mixed Model 3 examining socio-economic status as a
moderator of monolingual-bilingual differences SES in the full
Estimate SE t p
Intercept 1.940 0.080 24.200 0.000
IDS 0.051 0.066 0.781 0.436
HPP 0.189 0.063 2.990 0.004
Single Screen 0.202 0.064 3.170 0.003
Age -0.021 0.020 -1.050 0.292
Trial # -0.037 0.002 -19.500 0.000
NAE -0.018 0.051 -0.363 0.718
Bilingual 0.003 0.026 0.109 0.913
SES -0.001 0.004 -0.204 0.838
IDS * HPP 0.029 0.020 1.410 0.160
IDS * Single Screen -0.022 0.021 -1.040 0.296
Age * Trial # 0.001 0.000 4.280 0.000
IDS * Trial # -0.003 0.003 -0.949 0.343
IDS * Age 0.021 0.017 1.250 0.213
IDS * NAE 0.031 0.017 1.800 0.072
Age * NAE 0.003 0.007 0.443 0.657
IDS * Bilingual -0.007 0.020 -0.336 0.737
Age * Bilingual -0.002 0.008 -0.206 0.837
IDS * SES 0.004 0.003 1.220 0.222
Age * SES -0.001 0.001 -0.781 0.435
IDS * Age * NAE 0.012 0.005 2.230 0.026
IDS * Age * Bilingual -0.004 0.007 -0.597 0.550
IDS * Age * SES -0.001 0.001 -0.599 0.549
Exploratory analyses944
The relationship between NAE and IDS for bilingual infants who have945
some exposure to NAE.
In our second confirmatory analysis model (linear mixed model
2), we found that bilingual infants with more exposure to NAE showed stronger IDS947
preference. However, this initial analysis included a number of bilingual infants who were not
exposed to NAE at all (Figure 2). This raises the question of whether the relation between949
NAE and IDS preference may be primarily driven by the infants who were not learning NAE.
In the following analysis, we re-ran the pre-registered NAE-IDS model by restricting the951
model to infants who were exposed to NAE between 25% and 75% of the time. After952
pruning for non-convergence, the final model was:953
log lt trial type method +trial type trial num +age trial num+
trial type age exp nae+
(1 |subid)+
(1 |lab)+
(1 |item)
Based on 135 infants, the interaction between Trial Type and NAE exposure was still954
statistically significant (
= 0.00528,
= 0
= 0
0038). This result suggested that a
dose-response relationship between infants’ exposure to NAE and their preference for IDS956
over ADS was not driven by infants living in non-NAE contexts alone (see Table 7 for details
of the model).958
Table 7
Linear Mixed Model testing the effects of exposure to NAE-IDS
(restricted to bilingual infants living in NAE contexts).
Estimate SE t p
Intercept 1.910 0.168 11.400 0.000
IDS -0.211 0.132 -1.600 0.112
HPP 0.227 0.142 1.600 0.180
Single Screen 0.094 0.200 0.472 0.663
Age -0.009 0.036 -0.265 0.791
Trial # -0.041 0.006 -7.410 0.000
EXP_NAE -0.002 0.002 -0.783 0.434
IDS * HPP 0.016 0.063 0.260 0.795
IDS * Single Screen -0.115 0.081 -1.420 0.156
Age * Trial # 0.001 0.001 1.230 0.219
IDS * Trial # 0.016 0.008 1.990 0.048
IDS * Age 0.022 0.030 0.720 0.472
IDS * EXP_NAE 0.005 0.002 2.900 0.004
Age * EXP_NAE 0.000 0.001 -0.653 0.515
IDS * Age * EXP_NAE 0.000 0.001 0.054 0.957
General Discussion959
The current study was designed to better understand the effects of experience on the960
tuning of infants’ preference for infant-directed speech (IDS) compared to adult-directed961
speech (ADS). Bilingual infants’ language experience is split across input in two different962
languages, which are being acquired simultaneously. Bilinguals and monolinguals may thus963
differ in their preference for IDS. To explore this question, we used a collaborative, multi-lab
(N = 17 labs) approach to gather a large dataset of infants who were either 6-9- or965
12-15-months old and growing up bilingual (N = 333 bilingual infants in the final sample,966
and a lab-matched sample of N = 385 monolingual infants from the same communities).967
Data were collected as a companion project to ManyBabies 1 (ManyBabies Consortium, in968
press), which was limited to infants growing up monolingual. Overall, we found that969
bilingualism neither enhanced nor attenuated infants’ preference for IDS, with bilinguals970
showing a similar magnitude and developmental trajectory of IDS preference as monolinguals
from age 6 to 15 months.972
Although bilingual experience did not appear to moderate infants’ preference for IDS,973
we found striking evidence that experience hearing North-American English (NAE, the974
language of our stimuli) contributed to the magnitude of bilingual infants’ IDS preference.975
Bilinguals with greater exposure to NAE showed greater IDS preferences (when tested in976
NAE) than those who had less exposure to NAE. This relationship between NAE exposure977
and IDS preference was also observed in a subsample of bilingual infants all acquiring NAE,
but who varied in how much they were exposed to NAE relative to their other native979
language. These results converge with those from the larger ManyBabies 1 monolingual980
study, which reported that monolinguals acquiring NAE had a stronger preference for IDS981
than monolinguals acquiring another language. Importantly, our approach provides a more982
nuanced view of the relationship between NAE and IDS preference, and suggests that there
is a continuous dose effect of exposure on preference. Together, our findings have a number984
of implications for bilingual language acquisition during infancy. In the following, we will985
first discuss each of our research questions in turn, followed by limitations and implications986
of our study.987
Our first research question asked whether bilingualism affects infants’ attention to IDS
relative to ADS. We hypothesized that the complexity of the bilingual infant’s learning989
experience might lead to greater reliance on/preference for IDS, given that IDS may be990
viewed as an enhanced linguistic signal. However, this hypothesis was not confirmed. We991
observed a meta-analytic effect size in the full dataset for monolinguals of dz= 0.36 [CI =992
0.28 - 0.44] and for bilinguals of dz= 0.26 [CI: 0.09 - 0.43]. While monolinguals showed a993
numerically larger effect size, this difference was not statistically significant in either the994
meta-analyses or the mixed-effects linear models. Although small differences are still995
possible, our data generally support the conclusion that bilingual and monolingual infants996
show a similar preference for IDS over ADS. Specifically, both groups show a preference for997
IDS at 6-9 months of age, which gets stronger by 12-15 months.998
An additional part of our first research question asked whether bilinguals might show999
more variability than monolinguals in their IDS preference, beyond any differences in the1000
magnitude of the preference. We had reasoned that given their diversity of language1001
experiences, bilingual groups may have a higher heterogeneity in terms of their IDS1002
preference compared to monolingual groups (see also Orena & Polka, 2019, for a recent1003
experiment that observed this pattern). However, while both monolingual and bilingual1004
groups showed high variability, there were no reliable differences in variability observed1005
across groups. Thus, our results did not support the idea that bilingual infants show more1006
variable IDS preference than their monolingual peers.1007
Given that monolinguals and bilinguals can systematically differ in their1008
socio-economic status (SES), the third part of our first research question asked whether SES
might moderate bilingualism effects. We found mixed support for the role of SES in our1010
datasets. In our smaller dataset (matched dataset across labs), we found a statistically1011
significant interaction between age, SES, and IDS preference: 6-9-month-olds from higher1012
SES families showed stronger IDS preference than those from lower SES families, whereas1013
12-15-month-olds showed similar IDS preference regardless of SES. The direction of this1014
effect aligns with other research reporting that children from higher SES families generally1015
receive more language input and/or higher quality input (e.g., engaging in conversations1016
with more lexical diversity, complexity, and structural variations) than children from lower1017
SES families (Fernald, Marchman, & Weisleder, 2013; Hart & Risley, 1995; Hoff, 2006; Tal &
Arnon, 2018). Thus, this could suggest that infants from higher SES families may show1019
stronger IDS preference earlier in life as they hear more or higher quality IDS in their daily
lives. Further, this positive SES impact may be most beneficial to younger infants whose IDS
preference is still developing. However, given that in our larger (full) dataset SES was1022
unrelated to IDS preference in either 6-9- or 12-15-month-olds, this result might be spurious
and should be interpreted with caution.1024
Our second research question asked whether and how the amount of exposure to NAE
would affect bilingual infants’ listening preferences. Given that our stimuli were produced in
NAE, we expected that greater exposure to NAE would be linked to greater attention to1027
NAE IDS relative to NAE ADS. Indeed, ManyBabies 1 (ManyBabies Consortium, in press),
which was conducted concurrently with the current study, found that that monolinguals1029
acquiring NAE showed a stronger IDS preference than monolinguals not acquiring NAE.1030
However, in the ManyBabies 1 study, exposure to NAE-IDS was a binary variable – either1031
the infants heard only NAE or heard only a different language in their language1032
environments. In the current paper, bilinguals provide a more nuanced way to address this1033
question, as bilinguals’ exposure to NAE varied continuously between 25% and 75% (for1034
infants learning NAE as one of their native languages) or was near 0% (for infants learning1035
two non-NAE native languages). We found clear evidence for a positive dose-response1036
relationship between exposure to NAE and infants’ preference for NAE-IDS. This evidence –
that bilinguals with more exposure to NAE showed a stronger NAE-IDS preference – was1038
also present when focusing only on bilinguals who were learning NAE as one of their native
languages (i.e., those exposed to NAE 25-75% of the time). Importantly, we did not find a1040
similar effect of exposure to NAE on infants’ overall looking. This suggests that the effect of
NAE exposure on preference for IDS is a meaningful relationship, rather than an artefact1042
due to the stimuli being presented in NAE. Further studies with stimuli in other languages1043
would be necessary to solidify this conclusion.1044
As the first study to recruit and test bilingual infants at such a large scale and at so1045
many sites, we encountered several challenges (see also Byers-Heinlein et al., under review,1046
for a fuller discussion of challenges in planning and conducting ManyBabies 1). First, several
laboratories were not able to recruit the number of bilingual infants they had originally1048
planned. All labs committed to collecting a minimum of 16 bilingual infants per age group.1049
This ended up being unfeasible for some labs within the timeframe available (which was1050
more than a year), in some cases due to a high number of participants not meeting our strict
criterion for inclusion as bilingual. This undoubtedly highlights the challenges for labs in1052
recruiting bilingual infant samples, and moreover raises questions about how bilingualism1053
should be defined, and whether it should be treated as a continuous vs. categorical variable
(Anderson, Mak, & Bialystok, 2018; Bialystok, Luk, Peets, & Yang, 2018; Incera &1055
McLennan, 2018). Second, we had planned to explore the effect of different language pairs1056
on IDS preference. We had expected that some labs would be able to recruit relatively1057
homogeneous samples of infants (i.e., all learning the same language pair), but in the end1058
only one of 17 labs did so (another lab had planned to recruit a homogeneous sample but1059
deviated from this plan when it appeared unfeasible). Thus, we leave the question of the1060
effect of language pair on infants’ IDS preference an open issue to be followed up in future1061
studies. By and large, we believe that our large-scale approach to data collection may in the
future allow for the creation of homogeneous samples of infants tested at different1063
laboratories around the world. As such, large-scale and multi-site bilingual research projects
provide researchers with a powerful way to examine how the diversity and variability of1065
bilinguals impact their language and cognitive development.1066
Overall, our finding that bilinguals show similar preference for IDS as monolinguals1067
reinforces theoretical views that emphasize the similarities in attentional and learning1068
mechanisms across monolingual and bilingual infants (e.g., Curtin, Byers-Heinlein, & Werker,
2011). IDS appears to be a signal that enhances attention in infants from a variety of1070
language backgrounds. Yet, bilingual infants appear to be exquisitely fine-tuned to the1071
relative amount of input in each of their languages. It could have been the case that1072
language exposure has a threshold effect with any regular exposure to NAE enhancing1073
infants’ preference for NAE-IDS, marking it is a highly relevant speech signal. Instead, we1074
observed a graded effect such that the magnitude of bilingual infants’ preference varied1075
continuously with the amount of exposure to NAE. Just as bilingual infants’ relative1076
vocabulary size and early grammar skills in each language are linked to the amount of input
in that language (Hoff et al., 2012; Place & Hoff, 2011), the current study shows that the1078
amount of language input may also play an important role in other language acquisition1079
processes. Indeed, an intriguing but untested possibility is that different input-related1080
attentional biases (i.e., IDS preference) across bilinguals’ two languages explain important1081
variability in the early development of bilingual children’s vocabulary and grammar. Future
bilingual work can investigate the above possibility to further delineate the interplay between
infants’ language input, IDS preference, vocabulary, and grammar skills.1084
To conclude, the findings of the current study provide a more nuanced view of the1085
development of infants’ preference for IDS than prior studies have allowed. IDS preference1086
develops along a similar trajectory across infants from monolingual and bilingual1087
backgrounds. Importantly, by testing bilingual infants, our results revealed that this IDS1088
preference operates in a dose-response fashion, where the amount of exposure to NAE1089
positively moderated infants’ (NAE-) IDS preference in a continuous way. Our experience1090
highlights the challenges in recruiting and testing bilingual infants, but also reveals the1091
promise of large-scale collaborations for increasing sample sizes, and thus improving the1092
replicability and generalizability of key findings in infant research.1093
Author Contributions1094
Author contribution initials reflect authorship order. KBH, MCF, JG, MS contributed
to the study concept. KBH, MCF, JG, KK, CLW, MM, MS contributed to the study design.
KBH, CB contributed to the final protocol. KBH contributed to study documentation. KBH
contributed to study management. KBH, ASMT, AB, AB, SD, CTF, ACF, AG, JG, NGG,1098
JW contributed to data collection. KBH, ASMT, CB, MCF, JK contributed to data analysis.
KBH, CB, AB, MJC, CTF, MCF, JG, NGG, JKH, CLW, LS, MS contributed to the stage 1
manuscript. KBH, ASMT, CTF, MCF, JG, LS, MS contributed to the stage 2 manuscript.1102
Conflicts of Interest1103
The authors declare that there were no conflicts of interest with respect to the1104
authorship or the publication of this article.1105
Our manuscript was reviewed prior to data collection (; in1107
addition, we registered our instructions and materials prior to data collection1108
Data, materials, and online resources1110
All data and analytic code are available at1111 All materials are available via the1112
ManyBabies 1 monolingual Open Science Framework site at
Anderson, J., Mak, A., L.and Keyvani Chahi, & Bialystok, E. (2018). The language and1115
social background questionnaire: Assessing degree of bilingualism in a diverse1116
population. Behavior Research Methods,50 (1), 250–263.1117
Barr, D. J., Levy, R., Scheepers, C., & Tily, H. J. (2013). Random effects structure for1119
confirmatory hypothesis testing: Keep it maximal. Journal of Memory and Language,
68 (3), 255–278.
Bates, D., Kliegl, R., Vasishth, S., & Baayen, H. (2015a). Parsimonious mixed models.1122
Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015b). Fitting linear mixed-effects models
using lme4. Journal of Statistical Software, Articles,67 (1), 1–48.1125
Bialystok, E., Luk, G., & Kwan, E. (2005). Bilingualism, biliteracy, and learning to read:1127
Interactions among languages and writing systems. Scientific Studies of Reading,1128
9(1), 43–61.
Bialystok, E., Luk, G., Peets, K. F., & Yang, S. (2018). Receptive vocabulary differences in
monolingual and bilingual children. Bilingualism: Language and Cognition,13 (4),1131
Bosch, L., & Ramon-Casas, M. (2011). Variability in vowel production by bilingual speakers:
Can input properties hinder the early stabilization of contrastive categories? Journal
of Phonetics,39 (4), 514–526.1135
Bosch, L., & Sebastián-Gallés, N. (1997). Native-language recognition abilities in1137
4-month-old infants from monolingual and bilingual environments. Cognition,65 (1),1138
Bosch, L., & Sebastián-Gallés, N. (2001). Evidence of early language discrimination abilities
in infants from bilingual environments. Infancy,2(1), 29–49.1141
Brito, N., & Barr, R. (2014). Flexible memory retrieval in bilingual 6-month-old infants.1143
Developmental Psychobiology,56 (5), 1156–1163.
Brito, N., Sebastián-Gallés, N., & Barr, R. (2015). Differences in language exposure and its
effects on memory flexibility in monolingual, bilingual, and trilingual infants.1146
Bilingualism: Language and Cognition,18 (4), 670–682.1147
Byers-Heinlein, K. (2015). Methods for studying infant bilingualism. In J. W. Schwieter1149
(Ed.), The Cambridge Handbook of Bilingual Processing (pp. 133–154). Cambridge:1150
Cambridge University Press.
Byers-Heinlein, K., Bergmann, C., Davies, C., Frank, M. C., Hamlin, J. K., Kline, M., . ..1152
Soderstrom, M. (under review). Building a collaborative psychological science:1153
Lessons learned from manybabies 1. Canadian Psychology.1154
Byers-Heinlein, K., & Fennell, C. T. (2014). Perceptual narrowing in the context of increased
variation: Insights from bilingual infants. Developmental Psychobiology,56 (2),1157
Byers-Heinlein, K., Fennell, C. T., & Werker, J. F. (2013). The development of associative1159
word learning in monolingual and bilingual infants. Bilingualism: Language and1160
Cognition,16 (1), 198–205.
Byers-Heinlein, K., Schott, E., Gonzalez-Barrero, A. M., Brouillard, M., Dubé, D., Jardak,1162
A., . . . al. (2019). MAPLE: A multilingual approach to parent language estimates.1163
Bilingualism: Language and Cognition, 1–7.1164
Cattani, A., Abbot-Smith, K., Farag, R., Krott, A., Arreckx, F., Dennis, I., & Floccia, C.1166
(2014). How much exposure to English is necessary for a bilingual toddler to perform
like a monolingual peer in language tests? International Journal of Language &1168
Communication Disorders,49 (6), 649–671.
Choi, S., & Bowerman, M. (1991). Learning to express motion events in English and Korean:
The influence of language-specific lexicalization patterns. Cognition,41 (1), 83–121.1171
Cooper, R. P., Abraham, J., Berman, S., & Staska, M. (1997). The development of infants’1173
preference for motherese. Infant Behavior and Development,20 (4), 477–488.1174
Cooper, R. P., & Aslin, R. N. (1990). Preference for infant-directed speech in the first month
after birth. Child Development,61 (5), 1584–1595. Retrieved from1177
Cooper, R. P., & Aslin, R. N. (1994). Developmental differences in infant attention to the1179
spectral properties of infant-directed speech. Child Development,65 (6), 1663–1677.1180
Csibra, G., Hernik, M., Mascaro, O., Tatone, D., & Lengyel, M. (2016). Statistical treatment
of looking-time data. Developmental Psychology,52 (4), 521–536.1183
Curtin, S., Byers-Heinlein, K., & Werker, J. E. (2011). Bilingual beginnings as a lens for1185
theory development: PRIMIR in focus. Journal of Phonetics,39 (4), 492–504.1186
Danielson, D. K., Seidl, A., Onishi, K. H., Alamian, G., & Cristia, A. (2014). The acoustic1188
properties of bilingual infant-directed speech. The Journal of the Acoustical Society1189
of America,135 (2), EL95–EL101.
DeAnda, S., Bosch, L., Poulin-Dubois, D., Zesiger, P., & Friend, M. (2016). The Language1191
Exposure Assessment Tool: Quantifying language exposure in infants and children.1192
Journal of Speech, Language, and Hearing Research,59 (6), 1346–1356.1193
Dunst, C., Gorman, E., & Hamby, D. (2012). Preference for infant-directed speech in1195
preverbal young children. Center for Early Literacy Learning,5(1), 1–13.1196
Farran, L. K., Lee, C.-C., Yoo, H., & Oller, D. K. (2016). Cross-cultural register differences
in infant-directed speech: An initial study. PLOS ONE,11 (3), 1–14.1198
Fennell, C. T., & Byers-Heinlein, K. (2014). You sound like Mommy: Bilingual and1200
monolingual infants learn words best from speakers typical of their language1201
environments. International Journal of Behavioral Development,38 (4), 309–316.1202
Fennell, C. T., Byers-Heinlein, K., & Werker, J. F. (2007). Using speech sounds to guide1204
word learning: The case of bilingual infants. Child Development,78 (5), 1510–1525.1205
Fernald, A. (1985). Four-month-old infants prefer to listen to motherese. Infant Behavior1207
and Development,8(2), 181–195.1208
Fernald, A., Marchman, V. A., & Weisleder, A. (2013). SES differences in language1210
processing skill and vocabulary are evident at 18 months. Developmental Science,1211
16 (2), 234–248.
Fernald, A., Taeschner, T., Dunn, J., Papousek, M., Boysson-Bardies, B. de, & Fukui, I.1213
(1989). A cross-language study of prosodic modifications in mothers’ and fathers’1214
speech to preverbal infants. Journal of Child Language,16 (3), 477–501.1215
Frank, M. C., Bergelson, E., Bergmann, C., Cristia, A., Floccia, C., Gervain, J., ...1217
Yurovsky, D. (2017). A collaborative approach to infant research: Promoting1218
reproducibility, best practices, and theory-building. Infancy,22 (4), 421–435.1219
García-Sierra, A., Ramírez-Esparza, N., & Kuhl, P. K. (2016). Relationships between1221
quantity of language input and brain responses in bilingual and monolingual infants,
110, 1–17.
Glenn, S. M., & Cunningham, C. C. (1983). What do babies listen to most? A1224
developmental study of auditory preferences in nonhandicapped infants and infants1225
with Down’s syndrome. Developmental Psychology,19 (3), 332–337.1226
Graf Estes, K., & Hay, J. F. (2015). Flexibility in bilingual infants’ word learning. Child1228
Development,86 (5), 1371–1385.
Graf Estes, K., & Hurley, K. (2013). Infant-directed prosody helps infants map sounds to1230
meanings. Infancy,18 (5), 797–824.
Hart, B., & Risley, T. R. (1995). Meaningful differences in the everyday experience of young
american children. Baltimore: Paul H. Brookes Publishing.1233
Hayashi, A., Tamekawa, Y., & Kiritani, S. (2001). Developmental change in auditory1234
preferences for speech stimuli in Japanese infants. Journal of Speech, Language, and1235
Hearing Research,44 (6), 1189–1200.
Higgins, J. P., & Thompson, S. G. (2002). Quantifying heterogeneity in a meta-analysis.1237
Statistics in Medicine,21 (11), 1539–1558.
Higgins, J. P., Thompson, S. G., Deeks, J. J., & Altman, D. G. (2003). Measuring1239
inconsistency in meta-analyses. BMJ,327 (7414), 557–560.1240
Hoff, E. (2006). How social contexts support and shape language development.1242
Developmental Review,26 (1), 55–88.1243
Hoff, E., Core, C., Place, S., Rumiche, R., Señor, M., & Parra, M. (2012). Dual language1245
exposure and early bilingual development. Journal of Child Language,39 (1), 1–27.1246
Hudon, T. M., Fennell, C. T., & Hoftyzer, M. (2013). Quality not quantity of television1248
viewing is associated with bilingual toddlers’ vocabulary scores. Infant Behavior and
Development,36 (2), 245–254.1250
Incera, S., & McLennan, C. T. (2018). The time course of within and between-language1252
interference in bilinguals. International Journal of Bilingualism,22 (1), 88–99.1253
Karzon, R. G. (1985). Discrimination of polysyllabic sequences by one- to four-month-old1255
infants. Journal of Experimental Child Psychology,39 (2), 326–342.1256
Kemler Nelson, D. G., Jusczyk, P. W., Mandel, D. R., Myers, J., Turk, A., & Gerken, L.1258
(1995). The head-turn preference procedure for testing auditory perception. Infant1259
Behavior and Development,18 (1), 111–116.1260
Kitamura, C., & Lam, C. (2009). Age-specific preferences for infant-directed affective intent.
Infancy,14 (1), 77–100.
Kitamura, C., Thanavishuth, C., Burnham, D., & Luksaneeyanawin, S. (2001). Universality
and specificity in infant-directed speech: Pitch modifications as a function of infant1265
age and sex in a tonal and non-tonal language. Infant Behavior and Development,1266
24 (4), 372–392.
Klein, R. A., Ratliff, K. A., Vianello, M., Adams, R. B., Bahník, Š., Bernstein, M. J., ...1268
Nosek, B. A. (2014). Investigating variation in replicability. Social Psychology,45 (3),
Kovács, Á. M., & Mehler, J. (2009a). Cognitive gains in 7-month-old bilingual infants.1271
Proceedings of the National Academy of Sciences,106 (16), 6556.1272
Kovács, Á. M., & Mehler, J. (2009b). Flexible learning of multiple speech structures in1274
bilingual infants. Science,325 (5940), 611.
Kuznetsova, A., Brockhoff, P., & Christensen, R. (2016). Package ’ImerTest’.1276
Lewkowicz, D. J., & Ghazanfar, A. A. (2006). The decline of cross-species intersensory1277
perception in human infants. Proceedings of the National Academy of Sciences,1278
103 (17), 6771–6774.
Liu, L., & Kager, R. (2015). Bilingual exposure influences infant VOT perception. Infant1280
Behavior and Development,38, 27–36.1281
Liu, L., & Kager, R. (2017a). Is mommy talking to daddy or to me? Exploring parental1283
estimates of child language exposure using the Multilingual Infant Language1284
Questionnaire. International Journal of Multilingualism,14 (4), 366–377.1285
Liu, L., & Kager, R. (2017b). Perception of tones by bilingual infants learning non-tone1287
languages. Bilingualism: Language and Cognition,20 (3), 561–575.1288
Ma, W., Golinkoff, R. M., Houston, D. M., & Hirsh-Pasek, K. (2011). Word learning in1290
infant- and adult-directed speech. Language Learning and Development,7(3),1291
Maas, C. J. M., & Hox, J. J. (2005). Sufficient sample sizes for multilevel modeling.1293
Methodology: European Journal of Research Methods for the Behavioral and Social1294
Sciences,1(3), 86–92.
ManyBabies Consortium. (in press). Quantifying sources of variability in infancy research1296
using the infant-directed speech preference. Advances in Methods and Practices in1297
Psychological Science. Pre-Registered Report Accepted for Publication.1298
Mattock, K., Polka, L., Rvachew, S., & Krehm, M. (2010). The first steps in word learning1299
are easier when the shoes fit: Comparing monolingual and bilingual infants.1300
Developmental Science,13, 229–43.
Matuschek, H., Kliegl, R., Vasishth, S., Baayen, H., & Bates, D. (2017). Balancing Type I1302
error and power in linear mixed models, 94, 305–315.1303
Maurer, D., & Werker, J. F. (2014). Perceptual narrowing during infancy: A comparison of
language and faces. Developmental Psychobiology,56 (2), 154–178.1306
MetaLab. (2017). Interactive tools for community-augmented meta-analysis, , power1308
analysis, and experimental planning in language acquisition research. Retrieved from
Molnar, M., Carreiras, M., & Gervain, J. (2016). Language dominance shapes non-linguistic
rhythmic grouping in bilinguals. Cognition,152, 150–159.1312
Nakagawa, S., Poulin, R., Mengersen, K., Reinhold, K., Engqvist, L., Lagisz, M., & Senior,1314
A. M. (2015). Meta-analysis of variation: Ecological and evolutionary applications1315
and beyond. Methods in Ecology and Evolution,6(2), 143–152.1316
Newman, R. S., & Hussain, I. (2006). Changes in preference for infant-directed speech in low
and moderate noise by 4.5- to 13-month-olds. Infancy,10 (1), 61–76.1319
Open Science Collaboration. (2015). Estimating the reproducibility of psychological science.
Science,349 (6251).
Orena, A. J., Byers-Heinlein, K., & Polka, L. (2019). What do bilingual infants actually1323
hear? Evaluating measures of language input to bilingual-learning 10-month-olds.1324
Developmental Science,n/a(n/a), e12901.
Orena, A. J., & Polka, L. (2019). Monolingual and bilingual infants’ word segmentation1326
abilities in an inter-mixed dual-language task. Infancy,24 (5), 718–737.1327
Patihis, L., Oh, J. S., & Mogilner, T. (2015). Phoneme discrimination of an unrelated1329
language: Evidence for a narrow transfer but not a broad-based bilingual advantage.1330
International Journal of Bilingualism,19 (1), 3–16.1331
Pegg, J. E., Werker, J. F., & McLeod, P. J. (1992). Preference for infant-directed over1333
adult-directed speech: Evidence from 7-week-old infants. Infant Behavior and1334
Development,15 (3), 325–345.1335
Petitto, L. A., Berens, M. S., Kovelman, I., Dubins, M. H., Jasinska, K., & Shalinsky, M.1337
(2012). The “Perceptual Wedge Hypothesis” as the basis for bilingual babies’ phonetic
processing advantage: New insights from fNIRS brain imaging. Brain and Language,
121 (2), 130–143.
Place, S., & Hoff, E. (2011). Properties of dual language exposure that influence 2-year-olds’
bilingual proficiency. Child Development,82 (6), 1834–1849.1342
Polka, L., & Sundara, M. (2012). Word segmentation in monolingual infants acquiring1344
Canadian English and Canadian French: Native language, cross-dialect, and1345
cross-language comparisons. Infancy,17 (2), 198–232.1346
Pye, C. (1986). Quiché Mayan speech to children. Journal of Child Language,13 (1), 85–100.
Ramírez, N. F., Ramírez, R. R., Clarke, M., Taulu, S., & Kuhl, P. K. (2017). Speech1350
discrimination in 11-month-old bilingual and monolingual infants: A1351
magnetoencephalography study. Developmental Science,20 (1), e12427.1352
Ramírez-Esparza, N., García-Sierra, A., & Kuhl, P. K. (2014). Look who’s talking: Speech1354
style and social context in language input to infants are linked to concurrent and1355
future speech development. Developmental Science,17 (6), 880–891.1356
Ramírez-Esparza, N., García-Sierra, A., & Kuhl, P. K. (2017). The impact of early social1358
interactions on later language development in spanish–english bilingual infants. Child
Development,88 (4), 1216–1234.
Robertson, S., von Hapsburg, D., & Hay, J. S. (2013). The effect of hearing loss on the1361
perception of infant- and adult-directed speech. Journal of Speech, Language, and1362
Hearing Research,56 (4), 1108–1119.1363
Santesso, D. L., Schmidt, L. A., & Trainor, L. J. (2007). Frontal brain electrical activity1365
(EEG) and heart rate in response to affective infant-directed (ID) speech in1366
9-month-old infants. Brain and Cognition,65 (1), 14–21.1367
Schreiner, M. S., Altvater-Mackensen, N., & Mani, N. (2016). Early word segmentation in1369
naturalistic environments: Limited effects of speech register. Infancy,21 (5), 625–647.
Sebastián-Gallés, N., Albareda-Castellot, B., Weikum, W. M., & Werker, J. F. (2012). A1372
bilingual advantage in visual language discrimination in infancy. Psychological1373
Science,23 (9), 994–999.
Sebastián-Gallés, N., & Bosch, L. (2002). Building phonotactic knowledge in bilinguals: Role
of early exposure. Journal of Experimental Psychology: Human Perception and1376
Performance,28 (4), 974–989.
Senior, A. M., Gosby, A. K., Lu, J., Simpson, S. J., & Raubenheimer, D. (2016).1378
Meta-analysis of variance: An illustration comparing the effects of two dietary1379
interventions on variability in weight. Evolution, Medicine, and Public Health,1380
2016 (1), 244–255.
Shute, B., & Wheldall, K. (1999). Fundamental frequency and temporal modifications in the
speech of British fathers to their children. Educational Psychology,19 (2), 221–233.1383
Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2011). False-positive psychology:1385
Undisclosed flexibility in data collection and analysis allows presenting anything as1386
significant. Psychological Science,22 (11), 1359–1366.1387
Singh, L., Fu, C. S. L., Rahman, A. A., Hameed, W. B., Sanmugam, S., Agarwal, P., . . .1389
Team, the G. R. (2015). Back to basics: A bilingual advantage in infant visual1390
habituation. Child Development,86 (1), 294–302.
Singh, L., Morgan, J. L., & Best, C. T. (2002). Infants’ listening preferences: Baby talk or1392
happy talk? Infancy,3(3), 365–394.
Singh, L., Nestor, S., Parikh, C., & Yull, A. (2009). Influences of infant-directed speech on1394
early word recognition. Infancy,14 (6), 654–666.1395
Slobin, D. I. (Ed.). (1985). The crosslinguistic study of language acquisition (theoretical1397
issues) (1st ed., Vol. 2). New York: Psychology Press.1398
Soderstrom, M. (2007). Beyond babytalk: Re-evaluating the nature and content of speech1400
input to preverbal infants. Developmental Review,27 (4), 501–532.1401
Sundara, M., & Scutellaro, A. (2011). Rhythmic distance between languages affects the1403
development of speech perception in bilingual infants. Journal of Phonetics,39 (4),1404
Tal, S., & Arnon, I. (2018). SES effects on the use of variation sets in child-directed speech.
Journal of Child Language,45 (6), 1423–1438.1407
Tardif, T. (1996). Nouns are not always learned before verbs: Evidence from Mandarin1409
speakers’ early vocabularies. Developmental Psychology,32 (3), 492–504.1410
Tardif, T., Shatz, M., & Naigles, L. (1997). Caregiver speech and children’s use of nouns1412
versus verbs: A comparison of English, Italian, and Mandarin. Journal of Child1413
Language,24 (3), 535–565.
Thiessen, E. D., Hill, E. A., & Saffran, J. R. (2005). Infant-directed speech facilitates word1415
segmentation. Infancy,7(1), 53–71.
Trainor, L. J., & Desjardins, R. N. (2002). Pitch characteristics of infant-directed speech1417
affect infants’ ability to discriminate vowels. Psychonomic Bul letin & Review,9(2),1418
Viechtbauer, W. (2010). Conducting meta-analyses in r with the metafor package. Journal1420
of Statistical Software, Articles,36 (3), 1–48.
Werker, J. F., & Byers-Heinlein, K. (2008). Bilingualism in infancy: First steps in perception
and comprehension. Trends in Cognitive Sciences,12 (4), 144–151.1423
Werker, J. F., & McLeod, P. J. (1989). Infant preference for both male and female1425
infant-directed talk: A developmental study of attentional and affective1426
responsiveness. Canadian Journal of Psychology/Revue Canadienne de Psychologie,1427
43 (2), 230–246.
Werker, J. F., Pegg, J. E., & McLeod, P. J. (1994). A cross-language investigation of infant
preference for infant-directed communication. Infant Behavior and Development,1430
17 (3), 323–333.
Werker, J. F., & Tees, R. C. (1984). Cross-language speech perception: Evidence for1432
perceptual reorganization during the first year of life. Infant Behavior and1433
Development,7(1), 49–63.1434
Table A1
Number of monolingual and bilingual infants in each gender group by lab which met
infant-level inclusion criteria.
lab monolingual
babylabbrookes 18 12 14 20
babylabkingswood 11 19 9 15
babylabparisdescartes1 7 9 5 6
babylabprinceton 1 0 10 5
bllumanitoba 18 24 9 6
cdcceu 8 5 8 6
infantcogubc 8 3 7 3
infantstudiesubc 8 12 9 6
irlconcordia 15 20 16 18
isplabmcgill 5 6 8 8
langlabucla 1 2 5 4
ldlottawa 16 12 14 11
lllliv 17 17 4 9
lscppsl 7 7 7 9
nusinfantlanguagecentre 8 12 24 14
weltentdeckerzurich 14 16 16 12
wsigoettingen 17 29 5 11
Table A2
Number of bilingual infants per unique language pairs
language_pairs n
albanian ; non_nae_english 1
albanian ; swissgerman 1
arabic ; french 5
arabic ; german 1
arabic ; nae_english 2
arabic ; non_nae_english 2
armenian ; french 1
bahasa ; non_nae_english 1
belizean creole ; nae_english 1
bengali ; non_nae_english 1
bosnian ; non_nae_english 1
bulgarian ; german 1
cantonese ; german 1
cantonese ; nae_english 14
cantonese ; non_nae_english 2
dutch ; french 1
farsi ; non_nae_english 2
finnish ; german 1
finnish ; swissgerman 1
french ; georgian 1
french ; german 2
french ; hungarian 2
french ; italian 4
french ; korean 1
Table A2
Number of bilingual infants per unique language pairs (continued)
language_pairs n
french ; lebanese 1
french ; mandarin 1
french ; nae_english 64
french ; non_nae_english 9
french ; persian 1
french ; polish 1
french ; portuguese 2
french ; romanian 1
french ; russian 1
french ; spanish 6
french ; swissgerman 5
french.; kabyle 1
german ; hungarian 1
german ; kurdish 1
german ; lithuanian 1
german ; nae_english 5
german ; non_nae_english 9
german ; polish 2
german ; russian 2
greek ; non_nae_english 2
greek ; swissgerman 1
hebrew ; hungarian 3
hebrew ; nae_english 3
hindi ; non_nae_english 1
Table A2
Number of bilingual infants per unique language pairs (continued)
language_pairs n
hungarian ; italian 1
hungarian ; nae_english 1
hungarian ; non_nae_english 4
hungarian ; russian 2
hungarian ; spanish 1
indonesian ; nae_english 1
indonesian ; non_nae_english 1
italian ; nae_english 1
italian ; non_nae_english 2
italian ; swissgerman 3
japanese ; non_nae_english 3
khmer ; non_nae_english 1
korean ; nae_english 2
malayalam ; nae_english 1
mandarin ; nae_english 7
mandarin ; non_nae_english 44
nae_english ; persian 1
nae_english ; polish 1
nae_english ; punjabi 3
nae_english ; russian 3
nae_english ; spanish 17
nae_english ; swedish 2
nae_english ; swissgerman 1
nae_english ; tagalog 2
Table A2
Number of bilingual infants per unique language pairs (continued)
language_pairs n
nae_english ; telugu 1
nae_english ; urdu 1
nepali ; non_nae_english 1
non_nae_english ; patois 1
non_nae_english ; polish 7
non_nae_english ; portuguese 7
non_nae_english ; punjabi 1
non_nae_english ; russian 1
non_nae_english ; slovenian 1
non_nae_english ; spanish 7
non_nae_english ; swissgerman 5
non_nae_english ; tagalog 2
non_nae_english ; tamil 1
non_nae_english ; turkish 1
non_nae_english ; ukrainean 1
non_nae_english ; urdu 1
non_nae_english ; vietnamese 1
non_nae_english ; welsh 2
non_nae_english ; wu 1
portuguese ; swissgerman 1
romansh ; swissgerman 1
serbian ; swissgerman 1
slowenian ; swissgerman 1
spanish ; swissgerman 6
Table A2
Number of bilingual infants per unique language pairs (continued)
language_pairs n
swissgerman ; turkish 1
... Nine of the 11 participating labs were also participating in two prior multi-lab collaborative studies (ManyBabies 1 study and/or ManyBabies 1 Bilingual study) investigating infants' preference for infant-directed speech (Byers-Heinlein, Tsui, et al., 2020;ManyBabies Consortium, 2020). The current study emerged out of the unique opportunity afforded by a significant number of labs with a bilingual population coming together to run the Manybabies 1 Bilingual study, and the desire to make optimal use of these resources. ...
... This study forms part of a groundswell of large-scale, multi-lab initiatives all working towards the common goal of investigating generalizability and replicability of core findings in infant cognition (c.f., ManyBabies Consortium, 2020;Byers-Heinlein, Bergmann, et al., 2020;Byers-Heinlein, Tsui, et al., 2020). Sampling 322 infants distributed across 8 countries and 3 continents, this study provides confirmatory evidence for the replicability and generalizability of past evidence for infants' sensitivity to gaze cues. ...
Determining the meanings of words requires language learners to attend to what other people say. However, it behooves a young language learner to simultaneously encode relevant non‐verbal cues, for example, by following the direction of their eye gaze. Sensitivity to cues such as eye gaze might be particularly important for bilingual infants, as they encounter less consistency between words and objects than monolingual infants, and do not always have access to the same word‐learning heuristics (e.g., mutual exclusivity). In a preregistered study, we tested the hypothesis that bilingual experience would lead to a more pronounced ability to follow another's gaze. We used a gaze‐following paradigm developed by Senju and Csibra (Current Biology, 18, 2008, 668) to test a total of 93 6‐ to 9‐month‐old and 229 12‐ to 15‐month‐old monolingual and bilingual infants, in 11 laboratories located in 8 countries. Monolingual and bilingual infants showed similar gaze‐following abilities, and both groups showed age‐related improvements in speed, accuracy, frequency, and duration of fixations to congruent objects. Unexpectedly, bilinguals tended to make more frequent fixations to on‐screen objects, whether or not they were cued by the actor. These results suggest that gaze sensitivity is a fundamental aspect of development that is robust to variation in language exposure.
Determining the meanings of words requires language learners to attend to what other people say. However, it behooves a young language learner to simultaneously attend to what other people attend to, for example, by following the direction of their eye gaze. Sensitivity to cues such as eye gaze might be particularly important for bilingual infants, as they encounter less consistency between words and objects than monolinguals, and do not always have access to the same word learning heuristics (e.g., mutual exclusivity). In a pre-registered study, we tested the hypothesis that bilingual experience would lead to a more pronounced ability to follow another’s gaze. We used the gaze-following paradigm developed by Senju & Csibra (2008) to test a total of 93 6–9 month-old and 229 12–15 month-old monolingual and bilingual infants, in 11 labs located in 8 countries. Monolingual and bilingual infants showed similar gaze-following abilities, and both groups showed age-related improvements in speed, accuracy, frequency and duration of fixations to congruent objects. Unexpectedly, bilinguals tended to make more frequent fixations to onscreen objects, whether or not they were cued by the actor. These results suggest that gaze sensitivity is a fundamental aspect of development that is robust to variation in language exposure.
Full-text available
Research examining the cognitive consequences of bilingualism has expanded rapidly in recent years and has revealed effects on aspects of cognition across the lifespan. However, these effects are difficult to find in studies investigating young adults. One problem is that there is no standard definition of bilingualism or means of evaluating degree of bilingualism in individual participants, making it difficult to directly compare the results of different studies. Here, we describe an instrument developed to assess degree of bilingualism for young adults who live in diverse but predominantly English-speaking communities. We demonstrate the reliability and validity of the instrument in analyses based on 408 participants. The relevant factors for describing degree of bilingualism are (1) the extent of non-English language proficiency and use at home, and (2) non-English language use socially. We then use the bilingualism scores obtained from the instrument to demonstrate their association with (1) performance on executive function tasks, and (2) to previous classifications of participants into categories of monolinguals and bilinguals.
The field of infancy research faces a difficult challenge: some questions require samples that are simply too large for any one lab to recruit and test. ManyBabies aims to address this problem by forming large-scale collaborations on key theoretical questions in developmental science, while promoting the uptake of Open Science practices. Here, we look back on the first project completed under the ManyBabies umbrella - ManyBabies 1 - which tested the development of infant-directed speech preference. Our goal is to share the lessons learned over the course of the project and to articulate our vision for the role of large-scale collaborations in the field. First, we consider the decisions made in scaling up experimental research for a collaboration involving 100+ researchers and 70+ labs. Next, we discuss successes and challenges over the course of the project, including: protocol design and implementation, data analysis, organizational structures and collaborative workflows, securing funding, and encouraging broad participation in the project. Finally, we discuss the benefits we see both in ongoing ManyBabies projects and in future large-scale collaborations in general, with a particular eye towards developing best practices and increasing growth and diversity in infancy research and psychological science in general. Throughout the paper, we include first-hand narrative experiences, in order to illustrate the perspectives of researchers playing different roles within the project. While this project focused on the unique challenges of infant research, many of the insights we gained can be applied to large-scale collaborations across the broader field of psychology.
Psychological scientists have become increasingly concerned with issues related to methodology and replicability, and infancy researchers in particular face specific challenges related to replicability: For example, high-powered studies are difficult to conduct, testing conditions vary across labs, and different labs have access to different infant populations. Addressing these concerns, we report on a large-scale, multisite study aimed at (a) assessing the overall replicability of a single theoretically important phenomenon and (b) examining methodological, cultural, and developmental moderators. We focus on infants’ preference for infant-directed speech (IDS) over adult-directed speech (ADS). Stimuli of mothers speaking to their infants and to an adult in North American English were created using seminaturalistic laboratory-based audio recordings. Infants’ relative preference for IDS and ADS was assessed across 67 laboratories in North America, Europe, Australia, and Asia using the three common methods for measuring infants’ discrimination (head-turn preference, central fixation, and eye tracking). The overall meta-analytic effect size (Cohen’s d) was 0.35, 95% confidence interval = [0.29, 0.42], which was reliably above zero but smaller than the meta-analytic mean computed from previous literature (0.67). The IDS preference was significantly stronger in older children, in those children for whom the stimuli matched their native language and dialect, and in data from labs using the head-turn preference procedure. Together, these findings replicate the IDS preference but suggest that its magnitude is modulated by development, native-language experience, and testing procedure.
The field of infancy research faces a difficult challenge: some questions require samples that are simply too large for any one lab to recruit and test. ManyBabies aims to address this problem by forming large-scale collaborations on key theoretical questions in developmental science, while promoting the uptake of Open Science practices. Here, we look back on the first project completed under the ManyBabies umbrella – ManyBabies 1 – which tested the development of infant-directed speech preference. Our goal is to share the lessons learned over the course of the project and to articulate our vision for the role of large-scale collaborations in the field. First, we consider the decisions made in scaling up experimental research for a collaboration involving 100+ researchers and 70+ labs. Next, we discuss successes and challenges over the course of the project, including: protocol design and implementation, data analysis, organizational structures and collaborative workflows, securing funding, and encouraging broad participation in the project. Finally, we discuss the benefits we see both in ongoing ManyBabies projects and in future large-scale collaborations in general, with a particular eye towards developing best practices and increasing growth and diversity in infancy research and psychological science in general. Throughout the paper, we include first-hand narrative experiences, in order to illustrate the perspectives of researchers playing different roles within the project. While this project focused on the unique challenges of infant research, many of the insights we gained can be applied to large-scale collaborations across the broader field of psychology.
Examining how bilingual infants experience their dual language input is important for understanding bilingual language acquisition. To assess these language experiences, researchers typically conduct language interviews with caregivers. However, little is known about the reliability of these parent reports in describing how bilingual children actually experience dual language input. Here, we explored the quantitative nature of dual language input to bilingual infants. Further, we described some of the heterogeneity of bilingual exposure in a sample of French‐English bilingual families. Participants were twenty‐one families with a 10‐month‐old infant residing in Montréal, Canada. First, we conducted language interviews with the caregivers. Then, each family completed three full‐day recordings at home using the LENA (Language Environment Analysis) recording system. Results showed that children's proportion exposure to each language was consistent across the two measurement approaches, indicating that parent reports are reliable for assessing a bilingual child's language experiences. Further exploratory analyses revealed three unique findings: (1) there can be considerable variability in the absolute amount of input among infants hearing the same proportion of input, (2) infants can hear different proportions of language input when considering infant‐directed versus overheard speech, (3) proportion of language input can vary by day, depending on who is caring for the infant. We conclude that collecting naturalistic recordings is complementary to parent‐report measures for assessing infant's language experiences and for establishing bilingual profiles.
Bilingual infants vary in when, how, and how often they hear each of their languages. Variables such as the particular languages of exposure, the community context, the onset of exposure, the amount of exposure, and socioeconomic status are crucial for describing any bilingual infant sample. Parent report is an effective approach for gathering data about infants’ language experience. However, its quality is highly dependent on how information is elicited. This paper introduces a Multilingual Approach to Parent Language Estimates (MAPLE). MAPLE promotes best practices for using structured interviews to reliably elicit information from parents on bilingual infants’ language background, with an emphasis on the challenging task of quantifying infants’ relative exposure to each language. We discuss sensitive issues that must be navigated in this process, including diversity in family characteristics and cultural values. Finally, we identify six systematic effects that can impact parent report, and strategies for minimizing their influence.
Previous studies show that young monolingual infants use language‐specific cues to segment words in their native language. Here, we asked whether 8 and 10‐month‐old infants (N = 84) have the capacity to segment words in an inter‐mixed bilingual context. Infants heard an English‐French mixed passage that contained one target word in each language, and were then tested on their recognition of the two target words. The English‐monolingual and French‐monolingual infants showed evidence of segmentation in their native language, but not in the other unfamiliar language. As a group, the English‐French bilingual infants segmented in both of their native languages. However, exploratory analyses suggest that exposure to language mixing may play a role in bilingual infants’ segmentation skills. Taken together, these results indicate a close relation between language experience and word segmentation skills.
Socio-economic status (SES) impacts the amount and type of input children hear in ways that have developmental consequences. Here, we examine the effect of SES on the use of variation sets (successive utterances with partial self-repetitions) in child-directed speech (CDS). Variation sets have been found to facilitate language learning, but have been studied only in higher-SES groups. Here, we examine their use in naturalistic speech in two languages (Hebrew and English) for both low- and high-SES caregivers. We find that variation sets are more frequent in the input of high-SES caregivers in both languages, indicating that SES also impacts structural properties of CDS.
The ideal of scientific progress is that we accumulate measurements and integrate these into theory, but recent discussion of replicability issues has cast doubt on whether psychological research conforms to this model. Developmental research—especially with infant participants—also has discipline-specific replicability challenges, including small samples and limited measurement methods. Inspired by collaborative replication efforts in cognitive and social psychology, we describe a proposal for assessing and promoting replicability in infancy research: large-scale, multi-laboratory replication efforts aiming for a more precise understanding of key developmental phenomena. The ManyBabies project, our instantiation of this proposal, will not only help us estimate how robust and replicable these phenomena are, but also gain new theoretical insights into how they vary across ages, linguistic communities, and measurement methods. This project has the potential for a variety of positive outcomes, including less-biased estimates of theoretically important effects, estimates of variability that can be used for later study planning, and a series of best-practices blueprints for future infancy research.