Content uploaded by Deborah Nichols
Author content
All content in this area was uploaded by Deborah Nichols on Aug 22, 2018
Content may be subject to copyright.
On-screen print: the role of captions as
a supplemental literacy tool
Deborah Linebarger and Jessica Taylor Piotrowski
Annenberg School for Communication, University of Pennsylvania, Philadelphia,
PA, USA
Charles R. Greenwood
Juniper Gardens Children’s Project, University of Kansas, USA
Children living in poverty are 1.3 times as likely as non-poor children to experience
reading difficulties and lack key oral experiences that contribute to early literacy
development. The purpose of this research was to study the effects of viewing
commercially available educational television with closed captions. Seventy second-
and third-grade economically disadvantaged children living in urban locations
participated in this experimental research design. Children were randomly assigned
to view videos with or without closed captions. Captions helped children recognise
and read more words, identify the meaning of those words, generate inferences
regarding programme content and transfer these skills to a normative code-related
skill task. Risk status moderated word recognition performance: those at risk
benefited from captions while those who were not at risk recognised more words
when captions were absent.
The majority of children identified with learning disabilities have difficulties learning to
read (Snow et al., 1998). At the fourth grade level, a higher percentage of Asian/Pacific
Islander (i.e. 42%) and European American students (i.e. 41%) scored at or above
Proficient on the National Assessment of Educational Progress reading assessment when
compared with their American Indian/Alaska Native (i.e. 18%), Hispanic (i.e. 16%) and
African American (i.e. 13%) peers (Lee, Grigg & Donahue, 2007). Moreover, children
living in poverty are 1.3 times as likely as non-poor children to experience learning
disabilities and developmental delays (Brooks-Gunn & Duncan, 1997). Poor children
often have multiple risk factors that contribute to reading difficulties, most importantly,
lack of appropriate home literacy experiences (i.e. opportunities for verbal interactions,
storytelling and early book reading; Feitelson & Goldstein, 1986; Harris & Smith, 1987;
Snow et al., 1998). In addition, children who come from non-English-speaking homes
often with limited English proficiency are also at a disadvantage in learning to read
English. Both children who are poor and children who are English language learners may
also have poorly educated parents, little income, live in neighbourhoods where most
families share similar characteristics and attend schools where the student body is
Journal of Research in Reading, ISSN 0141-0423 DOI: 10.1111/j.1467-9817.2009.01407.x
Volume 33, Issue 2, 2010, pp 148–167
rUnited Kingdom Literacy Association 2009. Published by Blackwell Publishing, 9600 Garsington Road,
Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA
Journal of
Research in Reading
Journal of
Research in Reading
predominantly low achieving (Brooks-Gunn & Duncan, 1997; Snow et al., 1998), placing
them at greater risk for later reading difficulties. By virtue of these characteristics, poor
children and children with limited English proficiency are more likely to experience
reading failure and to be identified to receive special education services.
While it has been well documented that poor families lack educational resources at
home (e.g. books and stimulating toys and materials; Zill, Moore, Smith, Stief & Coiro,
1995), over 98% of US homes have televisions (Statistical Abstracts, 2007) and 92% of
these homes also have video cassette recorders or DVD players (Statistical Abstracts,
2007). Children whose parents have less than or equivalent to a high school diploma
watch significantly more television than children whose parents have some college
1
(Roberts, Foehr & Rideout, 2005), spend less time with print materials than children in
families where parents have more education (i.e. 36 vs 45 minutes) and perform better on
early literacy tasks when they have a television in their bedroom (Linebarger &
Wainwright, 2007). Given the extensive availability of this resource, one must ask how
can it be harnessed to increase the number of home literacy experiences and, thereby,
at-risk children’s literacy outcomes?
Models of learning from television
Dual coding theory. Media stimuli are complementary, and, as such, the multiple
modalities used in presenting these stimuli strengthen children’s encoding and subsequent
storage and retrieval of specific media content (Linebarger, 2001; Mayer & Anderson,
1991; Najjar, 1995; Neuman, 1995; Paivio & Csapo, 1973). When duplicative
information is simultaneously presented via two modalities (i.e. audio and visual
content), it can enhance young children’s understanding of programme content by serving
to increase the number of cognitive paths that can be followed to retrieve the information
(Najjar, 1995; Paivio, 1975), leading to increased learning (e.g. Linebarger, Kosanic,
Greenwood & Doku, 2004; Neuman, 1995). However, if information across modalities is
dissimilar, then young children will engage in one of two behaviours: (1) switch between
the two streams of information with subsequent reductions in acquisition of content in
both modalities; or (2) choose one modality at the expense of the other. When programme
content across modalities is dissimilar, young children tend to default to (i.e. recall more
content) the visual channel of information at the expense of the auditory track, leading
researchers to argue that young children privilege the visual domain (i.e. the visual
superiority effect; Nelson, Reed & Walling, 1976; Paivio, Rogers & Smythe, 1968;
Whitehouse, Maybery & Durkin, 2006). Without a formal content analysis, it is difficult
to say exactly how much educational television content across audio and visual tracks is
redundant. However, studies suggest that young children do not suffer comprehension
decrements when content is presented both aurally and visually.
Travelling lens theory. As with print media, acquiring information from television is
a complex process that involves both attention to and comprehension of programme
stimuli. When interacting with televised stimuli, children must attend to important
or interesting aspects of programme content as well as integrate this content into
meaningful, comprehensible bits of information (Huston & Wright, 1989). Children
develop cognitive skills to process information in one medium (e.g. television) and
are able to use these skills when processing content found in other media forms
(e.g. books; Linebarger, 2006; Neuman, 1995). Recently, researchers (e.g. Kendeou
CAPTIONED TELEVISION AND LITERACY 149
rUnited Kingdom Literacy Association 2009
et al., 2005; Linebarger & Piotrowski, 2009) have begun to examine how skills
developed in one medium transfer to and support processing in other media. For
example, research has shown that comprehension of televised stories at age 6 predicts
comprehension of print stories at age 8 (Kendeou et al., 2005). Similarly, television
narratives have been found to support the development of preschoolers’ story skills
and subsequently aid in picture-book comprehension (Linebarger & Piotrowski, 2009).
Anderson and his colleagues (e.g. Anderson & Lorch, 1983; Anderson, Lorch, Field &
Sanders, 1981) conducted a series of studies to understand how young children’s attention
is elicited during television viewing. The work resulted in a model of attention in which
comprehension of the televised messages drives further attention. Huston and Wright
(Huston & Wright, 1989; Rice, Huston & Wright, 1982) elaborated this model of
attention, arguing that a child’s attentional decisions when viewing a television
programme are a function of stimulus features (i.e. formal features such as sounds and
unusual visual effects; Calvert, Huston & Wright, 1987; Campbell, Wright & Huston,
1987); comprehensibility of the content; and viewer disposition (Huston and Wright
labelled the model the ‘travelling lens’). The model predicts that stimuli perceived as
‘moderately novel, of intermediate complexity, integratable, somewhat regular, partially
ordered, and recognizable’ (Huston & Wright, 1989, p. 117) should elicit the greatest
amounts of interest and attention. Similarly, familiarity and habituation as well as
perceived incomprehensibility should lead to low interest and attention. With age and
viewing experience, children will continually move towards more cognitively challenging
television stimuli. Thus, those stimuli that were once considered incomprehensible will
eventually move into the child’s ‘focal lens of maximum interest . . . before they are
habituated and become old hat’ (Huston & Wright, 1989, p. 118).
Application to literacy
Dual coding theory and the ‘travelling lens’ model predict that children with few print
experiences and poor reading skills may find on-screen print overly challenging and thus
not attend to it, whereas fluent readers may ignore on-screen print because they have
habituated to it (Linebarger et al., 2004). Although it is possible that through repetition
poor readers will gain enough familiarity with on-screen print to benefit from such
exposure (i.e. the combination of print and television provides an opportunity for media
synergy to occur), initially such benefits are unlikely. Similarly, fluent readers are
unlikely to experience much benefit from on-screen print as they are already progressing
on a positive literacy trajectory. It is the group in the middle, emerging readers who are
not yet fluent, who should evidence the greatest literacy gains from exposure to on-screen
print because the content will be perceived as both interesting and cognitively
challenging and thus within their ‘travelling lens’ (Huston & Wright, 1989; Linebarger
et al., 2004). There is preliminary evidence supporting these predictions; emerging
readers were found to benefit the most from on-screen print when compared with both
poor readers and fluent readers (Linebarger et al., 2004).
On-screen print
Televised on-screen print can take different forms including: (1) print placed on-screen as
part of a curriculum strategy in educational television (e.g. Between the Lions); (2)
captions; and (3) subtitling. Captioning and subtitling are superficially similar in that both
add words to a picture. Captions are intended for deaf or hard-of-hearing viewers, are
150 LINEBARGER, PIOTROWSKI and GREENWOOD
rUnited Kingdom Literacy Association 2009
usually broadcast in closed form (i.e. to see the captions, a viewer needs to activate the closed
caption option found in the television menu) and match the language featured in the audio.
Captions are usually placed in different areas ofthescreentoreflectthepositionofthespeaker
and they use symbols that indicate the presence of sound effects (e.g. ). Subtitles are meant
for hearing viewers; are usually positioned in the bottom centre of the screen; are ‘open’
(always visible to all viewers); ignore sound effects; and are a translation of the audio using a
different language. Although the two terms are used interchangeably intheUnitedKingdom,
the present study, conducted in the United States, is based on captions as defined above.
Impact of captions on word recognition abilities
When visible on-screen, captions provide an opportunity for children to read while they
view their favourite programmes. While viewing, children see words combined with
visual information and hear accompanying audio content that can help them match the
printed word form to its on-screen aural and visual referent. Koskinen and her
colleagues (Koskinen, Bowen, Gambrell, Jensema & Kane, 1997; Koskinen, Wilson,
Gambrell & Jensema, 1986; Neuman & Koskinen, 1992) proposed that adding captions
to programmes children already watch may subtly influence children’s knowledge and
understanding of print. In a longitudinal study of continuous caption use in the home,
children who viewed with captions scored significantly higher on normative tests of
word identification and passage comprehension when compared with non-caption
viewers (Koskinen et al., 1997). Linebarger (2001), using carefully controlled scripts
written at a reading level below the average child in the study (i.e. first-grade level),
found that second-grade children who viewed short clips with verbatim captions read
words found in the scripts more accurately when compared with children who viewed
the same clips without captions. In that study, the verbal narration was slower than
verbal narration typically found in commercial educational programming (i.e. 90 words
per minute [wpm] compared with a more typical 122 wpm in commercial educational
programming; Jensema, McCann & Ramsey, 1996). A number of other studies
examined whether educational television that uses on-screen print as part of a larger
curricular strategy (e.g. Between the Lions; not captions) could help children develop
literacy skills. In these studies, literacy skills’ improvements were found for children in
preschool through third grade (Linebarger et al., 2004; Linebarger & Wainwright, 2007;
Prince, Grace, Linebarger, Atkinson & Huffman, 2002; Uchikoshi, 2006). These studies
suggest that educational programmes with captions will support children’s recognition
of words used in these programmes.
Impact of closed captions on comprehension abilities
To comprehend televised texts effectively, children need to allocate a portion of their
cognitive capacity to actively distinguish, review and, in some cases, construct the
organising features (Baker & Brown, 1984). Shea (2000) argued that presenting content
via two modalities (i.e. aural and visual) helps children comprehend that content. Adding
a third modality (i.e. print), if it reinforces content found in the first two domains, has the
potential to further enhance comprehension. Alternatively, this additional content may
overwhelm children’s cognitive capacities, resulting in comprehension decrements.
Whether the third modality enhances or overwhelms cognitive processing depends on the
viewer, the viewer’s current skill set (e.g. can the viewer read and decode words?) and
general cognitive capacity available to simultaneously process this third stream of stimuli
CAPTIONED TELEVISION AND LITERACY 151
rUnited Kingdom Literacy Association 2009
as well as the nature of the tasks to be completed after viewing. Moreover, programme
comprehension requires that children retain relevant programme content (e.g. plot,
setting, characters, actions, goals) across programme segments, a skill significantly more
challenging than matching print with aural and visual stimuli.
In earlier research, the presence of on-screen print was associated with reductions in
young children’s comprehension of the televised content (Linebarger, 2001).
Identifying literal story elements involves summarising crucial story events as they
occurred in a televised text (i.e. when ‘X’ happened, what happened next) while
distinguishing inferential story elements involves identifying the organising features or
main idea of the story. Both types of comprehension are challenging for young children
in traditional contexts (i.e. book reading). Inferential identification usually appears in
the upper elementary grades (Baker & Brown, 1984; Ezell, Kohler, Jarzynka & Strain,
1992). Linebarger (2001) hypothesised that, because children were still struggling with
the mechanics of reading (i.e. they were beginning readers), the on-screen print
demanded and captured their attention, causing them to shift back and forth between the
storyline and the captioned text. Only one task could be done efficiently at a time
leaving children with fewer cognitive resources available to process the meaning of the
text both at a literal and an inferential level. Evidence that comprehension improves in
the presence of closed captions exists for children who are proficient readers (Griffin &
Dumestre, 1993). The research evidence suggests that children who view educational
television without captions will outperform their caption-viewing peers on comprehen-
sion tasks.
Impact of captions on phonological and phonemic awareness
Code-related literacy skills (i.e. phonological awareness, phonemic awareness) have been
linked to vocabulary size, narrative skill and print knowledge both in print contexts
(Dickinson, McCabe, Anastasopoulos, Peisner-Feinberg & Poe, 2003) and televised
contexts (Linebarger & Piotrowski, 2006). The relationship between code-related literacy
skills and vocabulary is derived from the development of children’s phonemic awareness.
That is, as children add more word forms to their lexicon, greater sensitivity to the subtle
phonemic variations in word forms (e.g. map vs mop) and how these variations are related
to different semantic representations is needed (Ouellette, 2006). As such, code-related
skills can be enhanced indirectly via contributions to general vocabulary knowledge.
Captions also directly impact code skills by providing the child with the opportunity to see
the printed form of a word, hear the aural representation and contextualise the meaning of
the word based on its televised representation (Sadoski & Paivio, 2001). These effects
should be strengthened with increased exposure. Furthermore, children who most benefit
tend to be those with various risk factors and, subsequently, difficulty acquiring literacy
skills (Anderson, Huston, Schmitt, Linebarger & Wright, 2001; Kulik & Kulik, 1991;
Linebarger et al., 2004). Thus, viewing educational television with captions should support
children’s phonological and phonemic awareness skills.
In summary, viewing television programmes with on-screen captions offers a
supplemental literacy tool that is both feasible and scalable. There is limited research
on the use of captions with beginning readers, particularly using commercially available
educational programming. The Linebarger (2001) study used carefully developed scripts
written at a first-grade level with more difficult words inserted. The current study
represents an important contribution to the literature, combining commercially available
152 LINEBARGER, PIOTROWSKI and GREENWOOD
rUnited Kingdom Literacy Association 2009
educational television and on-screen print especially as it relates to two under-served
populations: economically disadvantaged African American children and economically
disadvantaged Hispanic children who spoke Spanish as a first language (i.e. English
language learners). Acquisition of programme-specific vocabulary, improvements in
programme-specific comprehension and generalising these improvements to normative
measures of word recognition and phonemic awareness were tested. A child’s initial
reading ability was also tested to determine whether initial ability, used as an indicator of
risk status, moderated any of these results.
Method
Participants
Seventy second- and third-grade children attending four Title 1 public schools in two
large Midwestern cities in the United States whose participation was approved by their
parents were included. Forty-two were girls (M57.5 years) and 28 were boys (M57.8
years). Eighty-two per cent of the families reported incomes below US$30,000; 57%
were reading below a first-grade level and 79% of the children were reading below a
second-grade level as measured by their oral reading rates (M524.4 wpm; Oral Reading
Fluency Task; Good, Simmons, Kame’enui, Kaminski & Wallin, 2002); and five of the
children had an identified reading or learning disability as reported by their parents.
Children were recruited from a population of economically disadvantaged families. Fifty-
six per cent were African American children who spoke English as their first language
and 44% were Hispanic children learning English as a second language (i.e. English
language learners).
Design and intervention conditions
To address the hypotheses, an experimental design was used (Shadish, Cook & Campbell,
2001). Children were randomised into one of two conditions: intervention (n535) versus
control (n535). In the intervention group, children watched six videos with captions.
The control group watched the same set of six videos (in the same order) without
captions. See Table 1 for descriptive information by condition.
For those in the intervention condition, the on-screen closed captions were the original
captions created for each video by professional captioning agencies using parameters
established by the Media Access Group at WGBH (i.e. the Boston, Massachusetts public
television station) and the Federal Communication Commission (1999). These
parameters resulted in near-verbatim captions with a maximum presentation rate of
120–130 wpm.
2
When captions were edited to reduce the presentation rate, substitute
words or phrases were not more complicated than the original wording. With these
guidelines, captions for children’s programmes match the spoken narration or dialogue
about 84% of the time. We purposely used the original, commercially created captions to
examine whether commercially available captions could have the same impact as
carefully designed, researcher-derived captions could (e.g. Linebarger, 2001).
Setting and apparatus
Video episodes were presented to children using a 1300 TV/VCR combination unit. Children
were seated in groups of two to three, approximately 1–2 feet away from the screen and the
CAPTIONED TELEVISION AND LITERACY 153
rUnited Kingdom Literacy Association 2009
experimenter sat to the right and slightly behind the children. The viewing took place in
empty classrooms during the after-school programme hours at each of the schools.
Stimuli
Six 30-minute programmes airing during spring 1999 on the public broadcasting service
were chosen as stimuli. Curriculum documents and promotional materials developed by
each of the programmes indicated that these programmes had been designed for children
in our age group (i.e. second- and third-grade children). Programmes were viewed in the
following order: Arthur & Friends,Magic School Bus,Reading Rainbow,Wishbone,
Zoom and Kratts’Creatures.
Procedures
After parental consent was received, all children were pre-tested using the re-
searcher-developed and standardised instruments described below. Once completed, all
children began viewing stimuli associated with their assigned condition (i.e. with or
without captions) in small groups of two to three children. After viewing each video, all
Table 1. Demographic characteristics, frequency of reading and writing, and pre-test standardised reading
scores across all children and within group.
All No captions Captions F(1, 69) or
w
2
(70)
a
Child’s age (in years) (SDs) 7.63 7.57 (0.79) 7.70 (1.02) .38
Race/ethnicity w
2
5.06
African American 39 20 19
Hispanic
b
31 15 16
Gender w
2
5.00
Male 28 14 14
Female 42 21 21
Risk status w
2
5.42
At risk 14 6 8
Moderately at risk 26 13 13
Not at risk 30 16 14
Maternal education (SDs) 11.50 (2.14) 11.51 (1.72) 11.49 (2.52) .00
Child’s reading/writing frequency (%) .52
Rarely 31.4 25.7 37.1
Sometimes 31.5 37.2 25.8
Frequently 17.2 17.2 17.2
Daily 19.9 19.9 19.9
Normative pre-test literacy scores
Dolch words (SDs) 169.80 (46.49) 169.13 (48.76) 170.47 (44.80) .01
Decontextualised target words (SDs) 33.50 (30.02) 36.97 (31.36) 30.03 (28.64) .94
Oral reading fluency (SDs) 48.27 (32.72) 47.97 (35.01) 48.57 (30.76) .01
Nonsense word fluency (SDs) 53.55 (30.04) 57.06 (31.65) 50.04 (28.35) .96
Notes: Ranges for each variable: age 56–10 years; maternal education in years 57–16 years. Book reading and
writing frequency: rarely 5never to once a month; sometimes 5two times per month to once a week;
frequently 5three to four times per week; daily. Dolch words (out of 220) 59–219. Decontextualised target
words (out of 120) 50–112. Oral reading fluency 54–155 words per minute. Nonsense word fluency pre-test
score 58–141 letter–sounds per minute.
a
The chi-square tests and all F-tests were non-significant.
b
All children of Hispanic origin spoke Spanish as their first language and were categorised as English language
learners by their schools.
154 LINEBARGER, PIOTROWSKI and GREENWOOD
rUnited Kingdom Literacy Association 2009
children’s video-specific knowledge was tested, one-on-one, via a series of brief, video-
specific indices including: (1) word recognition of target words (i.e. words featured in the
video using criteria described below); (2) definition of target words; (3) one literal
comprehension question that required the child to identify the climax of the video; and
(4) one inferential comprehension question that required the child to describe the main
idea of the video. After all episodes were viewed, all children were again tested using the
original set of measures (or an alternate form, where available). All viewing and testing
sessions took place during the children’s after-school child care programme housed at
each of the schools involved (n54 schools).
Measures
The measurement strategy selected was based on several considerations including the
collection of information to describe the participants, sensitivity to immediate effects of
the intervention, the ability to detect change using standardised measures and robustness
to implementation quality.
Parent demographic and media access/use questionnaire. Parents were asked to provide
both demographic and media access/use information about their families and children.
Data used in these analyses included birth dates, gender, family size, income, ethnicity/
race, primary language in the home, whether or not the target child had an identified
disability, children’s use of media (i.e. computers, TV, print) in the home and the number
of books the child had available to him/her.
Video-specific literacy measures
Word recognition. All children were asked to read five words used in each video after
viewing the video. These words (i.e. target words) were selected based on the following
criteria. Each word needed to be: (1) a ‘content’ word (i.e. important to the video story
and not an incidental word like ‘because’, ‘through’ or ‘actually’); (2) repeated at least
four or five times in the video; (3) explained in the video either through a verbal
definition/description, through a visual demonstration or both (e.g. ‘humpback’ was
presented and defined by a scientist through a voice-over while viewers saw a swimming
and leaping humpback whale); and (4) at a second- to fourth-grade level. In our initial
selection of words, we used a book for authors to determine at what grade and in what
aspect of the curriculum a word was introduced (Mogilner, 1992). However, on closer
examination using another basic vocabulary text (i.e. the EDL Core Vocabularies in
Reading, Mathematics, Science, and Social Studies; Taylor et al., 1989), we determined
that our selected target words actually ranged in difficulty between first and seventh
grade. Because of the range of word grade levels, we weighted the words in the analyses
based on grade (e.g. Grade 2 word scores multiplied by 2, Grade 3 word scores multiplied
by 3 and Grade 4 word scores multiplied by 4). Project staff scored children’s
pronunciation of words for fluency (i.e. target word recognition; 4-point scale:
05inaccurate;15read with difficulty;25sounded out;35fluent; read with no
difficulty). Examiners were trained and practised until inter-examiner agreement on
fluency scores equalled 90% or higher.
Comprehension. Three different measures of video-specific comprehension were
administered to all children: target word comprehension, literal comprehension and
CAPTIONED TELEVISION AND LITERACY 155
rUnited Kingdom Literacy Association 2009
inferential comprehension. Given the number of videos each child needed to see, the
number of children involved in the study and the limited amount of time each child was
available to participate in this study, the comprehension questions needed to be brief.
After completing the word recognition task described above, children were asked to
provide a definition for each of the words (i.e. target word comprehension). In the event
that the child was unable to read the target word fluently, a clear pronunciation of the
word was provided before asking for the definition. Literal comprehension involved
identifying a critical or main story event that was visually and/or verbally described (e.g.
when x happened what happened next) while inferential comprehension required the
children to give the main idea of the programme.
Comprehension questions and definitions were both scored using a 3-point scale
(0 5no response;15partially correct answer;25accurate and complete answer). Two
independent raters scored 11% of the definitions. Krippendorff’s as (Hayes &
Krippendorff, 2007) were acceptable for target word comprehension ( 5.95), literal
comprehension ( 5.91) and inferential comprehension ( 5.86).
Normative literacy measures
Normative literacy measures were administered at both the pre-test and the post-test to all
participants.
General word recognition. First, children were asked to read a list containing 220 high-
frequency words (i.e. Dolch words) compiled by E. W. Dolch (1948). This list comprises
50–75% of the reading material encountered by American students, especially in
Kindergarten through third grade (Frey, Kress & Fountoukidis, 2000; May, 1998). We
included this list of words because all words except ‘upon’ were found multiple times
across the video stimuli. Next, children were asked to read a decontextualised target word
list that contained all 30 target words (i.e. five words from each of the six videos that
were selected based on the four considerations detailed above). As with the target word
video-specific assessments, children’s ability to read each word fluently was examined
(i.e. 4-point scale: 0 5inaccurate;15read with difficulty;25sounded out;35fluent;
read with no difficulty).
Normative code-related literacy skills. Two of five individually administered and timed
subtests from the Dynamic Indicators of Basic Early Literacy Skills (Good & Kaminski,
2002; Kaminski & Good, 1996) were used to measure children’s phonemic awareness
and fluency skills. Children were given a set of instructions, a practice item and then
asked to do as many of the particular subtest as they could for 1 minute. Each subtest had
20 equivalent forms available. Nonsense word fluency (NWF) consisted of made-up
words that follow the rules of English syllable structures and the alphabetic principle. In
every language there are restrictions about the patterning of phonemes (or sound
combinations) in words. Children were given a sheet of paper with nonsense words
printed across the page and were instructed to read each word or say as many of the
sounds as they could in 1 minute. The examiner gave credit for any correctly said sounds.
For example, a child would see ‘rij’ and say ‘/r/ /i/ /j/’ and receive 3 points or say ‘/r/’ and
receive 1 point. Oral reading fluency (ORF) consisted of children reading passages that
measure the child’s skill at recognising and reading words rapidly and accurately.
Interviewers recorded the number of words read accurately in 1 minute to determine the
child’s oral reading rate.
156 LINEBARGER, PIOTROWSKI and GREENWOOD
rUnited Kingdom Literacy Association 2009
Assignment to risk status
We used initial reading ability, identified by scores on the pre-test NWF assessment
(discussed above), as a marker of the child’s reading risk status. This score was used to
form three categories: at risk for poor reading outcomes in fourth grade, moderately at
risk and not at risk for reading difficulties in fourth grade using national benchmarks
empirically established by the creators of the measure (Good et al., 2002). Achieving 50
letter–sound correspondences by winter of second grade has been set as the benchmark
for prediction of successful later reading achievement while reading o30 letter–sounds
per minute is predictive of poor reading outcomes in fourth grade. At-risk performers
were those not making satisfactory progress towards learning to read at the start of the
investigation. Based on benchmarks for children this age, these children were in need of
intensive instructional support if they were to achieve desired reading outcomes in fourth
grade. Those whose scores reached the established benchmark were considered not at
risk. According to Good et al. (2002), parents and teachers can be confident that these
children are making adequate progress towards reading outcomes. The group of children
who were classified as moderately at risk had not achieved the benchmark indicating
adequate reading progress, yet they were not below the at-risk benchmark. No clear
prediction is available for these students; however, they may be moving off track in their
reading development (Good et al., 2002). Other research using a similar risk status
classification has resulted in more clearly distinguishing results of an on-screen print
intervention (Linebarger et al., 2004).
Results
Analysis of covariance (ANCOVA) models were used to evaluate group differences on
the post-test video-specific and normative measures. Between-subjects factors included
group (two levels: captions and non-captions), child’s risk status (three levels: at risk,
moderately at risk, not at risk) and child’s gender (two levels). Preliminary ANOVAs
examining each factor indicated that child’s gender was unrelated to any of the child
outcomes and was, therefore, dropped from the final models. We also examined whether
the children’s background (i.e. African American English-speaking or Hispanic English
language learners) impacted our results. Across the outcomes, the findings were either
that English language learners benefited from the caption intervention and English
speakers’ performance did not differ (i.e. target word, literal and inferential
comprehension); that both benefited from the intervention, but the effects were larger
for ELL children (i.e. target word recognition, Dolch words and NWF); or that there were
no benefits of the intervention for either group (i.e. ORF, decontextualised target word
recognition). Because our hypotheses were related to the efficacy of the delivered
intervention and the role that risk status played in that effectiveness, we controlled for the
child’s background in our analyses.
Several other covariates were used or constructed to control for potential third
variables. All analyses controlled for children’s scores on a particular outcome at pre-test
(e.g. pre-test NWF scores were controlled for when analysing post-NWF outcomes). For
analyses related to specific programme content, the pre-test target word recognition score
was used as a covariate. For analyses related to normative assessments, pre-test scores for
each normative measure were controlled. Finally, in all analyses, maternal and child
composites were used as covariates. The maternal composite, focusing on maternal
CAPTIONED TELEVISION AND LITERACY 157
rUnited Kingdom Literacy Association 2009
education, was constructed by z-score transforming and then summing reported maternal
years of education and mother’s highest degree. Mother’s education is a commonly used
covariate that has been found to relate to both child media use and literacy or academic
achievement outcomes (Anderson et al., 2001; Linebarger et al., 2004; Wright et al.,
2001). The child composite, focusing on frequency of child’s literacy-related behaviours,
was formed by z-score transforming and summing how often a child read books alone or
wrote letters, stories and other print-related activities.
The computed ANCOVA models allowed us to examine an overall or omnibus F-test
for the significance of the risk by group interaction effects. We also executed planned
comparisons examining group differences within each level of risk to determine which, if
any, of the comparisons were most involved with a significant omnibus interaction effect,
or, alternatively, if the omnibus was not significant, to evaluate whether any of the simple
main effects were significant. Corrections for experiment-wise error were performed
using modified Bonferroni adjustments of the alevel (i.e. reducing Type 1 error rates or
finding a significant difference when one does not exist; Jaccard, 1998). To estimate the
practical significance of the outcomes, effects sizes were computed (i.e. Cohen’s d;
Cohen, 1988). Table 2 reports all findings from the final ANCOVA analyses.
Impact of captions and video-specific content
Video-specific word recognition. We expected that children in the captions condition
would outperform children in the non-captions condition on the target word recognition
task and that these results would be moderated by risk status. Our results showed that
children who viewed programming with captions did not perform significantly differently
from their non-caption viewing peers on the target word recognition measure, although
trends were in favour of the captions group (no captions: M5150.64, SD 559.93;
captions: M5165.25, SD 556.85; d50.20). As predicted, we did find that risk status
moderated word recognition outcomes. Specifically, a significant two-way interaction
between group and risk status was present, F(2, 69) 54.327, p5.018, indicating that at-
risk (d50.61) and moderately at-risk (d50.84) children who viewed programmes with
captions did benefit from the intervention when compared with their counterparts who
viewed without captions while not-at-risk children who viewed without captions
outperformed their peers who viewed with captions (d5!0.45; see Figure 1). There
were no significant univariate tests indicating that all simple main effects of group were
contributing equally to the pattern of relationships identified in the interaction.
Video-specific comprehension. We hypothesised that children in the non-captions group
would outperform their peers in the captions group on the comprehension tasks and that
risk status would moderate these findings. Unexpectedly, we found that target word
comprehension and inferential comprehension scores were higher after viewing
programmes with captions versus viewing programmes without captions while literal
comprehension scores were unaffected by viewing condition. Further, risk status did not
moderate these findings.
Children who viewed with captions knew the meaning of more target words
(M5136.72, SD 547.09) when compared with children who viewed without captions
(M5113.22, SD 549.64), F(2, 69) 54.157, p5.046, d50.49. There were no
significant main effects or interactions for literal comprehension. Means for both groups
were nearly identical (captions: M59.46, SD 52.19; no captions: M59.77, SD 52.13;
158 LINEBARGER, PIOTROWSKI and GREENWOOD
rUnited Kingdom Literacy Association 2009
d5!0.14), suggesting that all children were able to adequately identify the critical story
event in each video. Finally, for inferential comprehension, there was a significant main
effect of group, F(1, 69) 54.023, p5.049, d50.48. Children who viewed with captions
were better able to provide the main idea of each programme (M511.05, SD 52.54)
compared with children who viewed without captions (M59.79, SD 52.72).
Table 2. Means, standard errors and ANCOVA results for group by risk status across all outcomes.
Outcome Non-captions Captions ANCOVA Fand (partial Z
2
)
M SE M SE Group (G) Risk (R) G "R
Target word recognition 1.101 (.018) 0.671 (.022) 4.327
*
(.126)
At risk 128.18 23.94 159.28 20.34
Moderately at risk 140.54 15.11 187.89 15.72
Not at risk 183.21
a
14.76 148.56
a
15.44
Group means 150.64 10.13 165.25 9.61
Target word comprehension 4.157
*
(.065) 0.408 (.013) 0.868 (.028)
At risk 109.88
b
19.84 157.89
b
26.85
Moderately at risk 115.36 12.52 124.81 13.02
Not at risk 114.41 12.22 127.47 12.79
Group means 110.69 9.43 131.26 8.85
Literal comprehension 0.352 (.006) 1.369 (.044) 1.006 (.032)
At risk 9.59 0.89 10.73 0.75
Moderately at risk 9.42 0.56 8.72 0.58
Not at risk 9.39 0.55 9.86 0.57
Group means 9.46 0.37 9.77 0.36
Inferential comprehension 4.023
*
(.063) 0.698 (.023) 0.805 (.026)
At risk 9.90
a
1.08 12.25
a
0.92
Moderately at risk 10.03 0.68 10.25 0.71
Not at risk 9.46 0.66 10.65 0.70
Group means 9.79 0.46 11.05 0.43
Dolch words 5.160
*
(.080) 0.017 (.001) 2.97
+
(.091)
At risk 164.98 8.80 178.98 8.84
Moderately at risk 160.70
b
5.80 185.91
b
6.28
Not at risk 174.43 5.51 172.38 5.86
Group means 166.70 3.87 179.09 3.87
Decontextualised word recognition 1.651 (.027) 0.052 (.002) 1.035 (.034)
At risk 34.72
a
5.49 45.26
a
5.02
Moderately at risk 39.32 3.49 42.82 3.79
Not at risk 40.70 3.50 39.34 3.55
Group means 38.25 2.33 42.47 2.35
Nonsense word fluency 4.83
*
(.076) 2.851
+
(.088) 0.263 (.009)
At risk 64.47 12.49 77.04 12.32
Moderately at risk 66.27 7.82 87.54 8.22
Not at risk 45.72 9.22 57.09 8.41
Group means 58.82 4.91 73.89 5.14
Oral reading fluency 0.190 (.003) 0.853 (.028) 0.104 (.004)
At risk 59.22 8.56 59.37 7.94
Moderately at risk 58.04 5.52 52.66 5.86
Not at risk 63.86 5.38 62.36 5.50
Group means 60.37 3.68 58.13 3.64
Notes: Group by risk status mean differences (i.e. means in each row) sharing the same superscript are either
marginally significant (
a
) or significantly different (
b
).
+
po.100;
*
po.05.
CAPTIONED TELEVISION AND LITERACY 159
rUnited Kingdom Literacy Association 2009
Transfer of video-specific content to normative literacy outcomes
General word recognition. We expected that viewing captions over time would support
children’s ability to transfer video-specific recognition of target words to decontextua-
lised word recognition tasks at the conclusion of the intervention when compared with
children watching the same programmes without captions. Further, we expected that risk
status would moderate these findings.
For the Dolch words tasks, we found a significant main effect of group,
F(1, 69) 55.160, p5.027, d50.54. Caption viewers were able to accurately and
fluently read more Dolch words (M5179.09, SD 522.90) when compared with non-
caption viewers (M5166.70, SD 522.90). There was also a marginally significant two-
way interaction between group and risk status, F(2, 69) 52.97, p5.059. At-risk
(d50.74) and moderately at-risk (d51.06) children who viewed programmes with
captions outperformed their counterparts who viewed without captions while not-at-risk
children who viewed without captions outperformed their peers who viewed with
captions (d5!0.06; see Figure 2). Univariate tests indicated that moderately at-risk
children were most involved in the significant result, F(1, 60) 57.43, p5.008. There
were no significant main effects or interactions for decontextualised target word
recognition, although the trend favoured caption viewers over non-caption viewers
(captions: M542.47, SD 513.90; no captions: M538.25, SD 513.78; d50.30).
Normative code-related skills. In line with the extant literature, we expected children in
the captions viewing condition to outperform their non-captions viewing peers on
normative code-related skill assessments, and for risk status to moderate this relationship.
For the NWF task, there was a significant main effect of group, F(1, 69) 54.830,
p5.032, d50.52. Caption viewers were able to read more nonsense words in 1 minute
(M573.89, SD 529.05) when compared with non-caption viewers (M558.82,
SD 529.05). There was also a marginally significant main effect of risk status,
F(1, 69) 52.851, p5.066. Moderately at-risk viewers (M576.91, SD 531.05) out-
performed at-risk viewers (M570.76, SD 555.05; d50.14) who, in turn, outperformed
not-at-risk viewers (M551.41, SD 540.37; d50.40). There were no significant main
0
50
100
150
200
At Risk Some Risk No Risk
Mean Target Word Recognition
No
Captions
Captions
Figure 1. Two-way interaction between risk status and group for target word recognition.
160 LINEBARGER, PIOTROWSKI and GREENWOOD
rUnited Kingdom Literacy Association 2009
effects or interactions for ORF. Means for both groups were nearly identical (captions:
M558.13 wpm, SD 521.53; no captions: M560.37, SD 521.77; d5!0.10).
Discussion
The majority of outcomes examined indicated that children who viewed with captions
outperformed their counterparts who viewed without captions. Risk status moderated the
relationship for target word recognition as well as Dolch words.
Impact of captions on word recognition
Dual coding theory predicts that the information delivered via more than one modality
(e.g. aural, video, print) should be better learned than information delivered via one
modality (Paivio & Csapo, 1973). We did find that word recognition scores for both
programme-specific content (i.e. target word recognition) and more generic content (i.e.
Dolch words) were higher for caption viewers when compared with non-caption viewers.
Programme-specific performance was moderated by risk status, indicating that effects
were limited to only at-risk and moderately at-risk children. This finding is explained by
the ‘travelling lens’ model. Specifically, children who have at least some familiarity with
print, but not so much that reading print is automatic or easy, would be most likely to
attend to the print due to its novelty and challenge. Earlier research found that children
most at risk and those not at risk did not benefit from educational television containing
on-screen print (Linebarger et al., 2004); however, the stimuli used in this earlier study
displayed only selected content in a print format and the children were younger (i.e.
Kindergarten and first grade). It is likely that by second and third grade, the children still
struggling with reading had enough print familiarity to find the presented content
moderately novel and challenging and thereby worthwhile to attend to. Conversely, those
children who had learned to read either found the content too easy and tuned it out (i.e.
not at risk) or they already possessed the skills measured in this study. Although
0
50
100
150
200
At Risk Some Risk No Risk
Mean Number of Dolch Words
No
Captions
Captions
Figure 2. Two-way interaction between risk status and group for mean number of Dolch words.
CAPTIONED TELEVISION AND LITERACY 161
rUnited Kingdom Literacy Association 2009
decontextualised target word recognition mean scores were higher for caption viewers at
the post-test, this difference was not significant.
In addition to displaying on-screen print tied to the aural narration, repetition of target
words probably played a role in learning to read these words. Target words repeated, on
average, 25 times per video and, when originally selected for this study, did not repeat in
any of the other videos children would see. This level of exposure was enough to help
children identify and read the words immediately after viewing, but not to retain or transfer
their knowledge of that word to a decontextualised reading task after a period of 10 days to
2 weeks. The improved performance for Dolch words, which did repeat in each video (i.e.
48.3% of the words, on average, in each video were comprised of Dolch words), suggests
that repeat exposure to on-screen print paired with visual and auditory inputs does help
children learn, retain and transfer knowledge of these words to a delayed reading task.
Impact of captions on comprehension scores
As with word recognition, dual coding theory predicts that comprehension of programme
content would be enhanced when that content was delivered in two modalities; however,
the effects of content delivered in three modalities is unclear. Linebarger (2001) found
that aural narration coupled with visual information supported children’s incidental and
inferential comprehension abilities (just two modalities) while on-screen print when
coupled with both aural narration and visual information (three modalities) supported
literal comprehension (i.e. when ‘x’ happened, what happened next). Given that
Linebarger’s (2001) sample was most similar to our sample, we hypothesised that
comprehension scores would be the highest for children viewing without captions
because only two modalities would be involved in the delivery of content and would,
therefore, not overload children’s limited cognitive resources. Despite these concerns, we
found that comprehension scores on two of the three comprehension tasks favoured
caption viewers over non-caption viewers (i.e. target word comprehension, inferential
comprehension). Further, the remaining outcome, literal comprehension scores, was
unaffected by the presence or absence of captions. Rather than overloading the child’s
cognitive capacity, it appears that the delivery of information via three modalities helped
the viewer focus on the episode content.
It may also be the case that children spent little time fixating on the on-screen print
(although word recognition scores would suggest that at least some attention was paid to
the print). In eye-tracking work with younger children, there were very few fixations to
print in both traditional stories (i.e. books; 4.5% of fixations to print; Evans & Saint-
Aubin, 2005; Justice, Skibbe, Canning & Lankford, 2005) and televised stories (Vaala,
Lapierre & Linebarger, 2009). By adulthood, 84% of fixations were to on-screen print
(Jensema, El Sharkawy, Danturthi, Burch & Hsu, 2000). Although not an exact
comparison, interpolating from these studies suggests that children in our sample likely
paid more attention than the younger children in Linebarger’s (2001) study but not so
much that they were unable to comprehend word meanings and content.
Impact of captions on code-related skills
Identification of nonsense words measured children’s knowledge of the alphabetic
principle and their understanding of English orthographic structure and was improved in
the presence of captions. The triple modality inputs for words (i.e. auditory, visual,
linguistic) that the captions-viewing group experienced helped children increase their
162 LINEBARGER, PIOTROWSKI and GREENWOOD
rUnited Kingdom Literacy Association 2009
stores of semantic and syntactic abilities, making them more fluent on this task (i.e. their
ability to identify sounds increased from 50 spm before the intervention to 66 spm after
the intervention). However, transferring this sound-specific fluency to more fluent reading
of generalised English-reading passages (i.e. ORF) did not occur. This is very close to our
findings reported here. Linebarger (2001) also found no impacts of captions on children’s
ORF scores. As with the decontextualised target word recognition task described above, it
is likely that viewing six different educational television programmes was not sufficient
to move the child’s general reading fluency abilities.
Although captions were not found to enhance children’s reading fluency, the findings
in favour of captions for supporting nonsense word identification are important. Fisch
(2000, 2004) discusses the importance of transfer of learning, or the ability to take
programme-specific content, mentally encode that content and then use it appropriately in
novel situations. These normative reading assessments measure children’s ability to
transfer the specific literacy content featured in the videos into more generalised literacy
skills. One of the most important educational goals for children is learning to transfer
specific skills to generalised contexts. Fisch (2004) suggested that this goal is best
supported through varied practice, or exposure to multiple examples in a variety of
formats. Educational television coupled with captions provided multiple and varied
formats that helped young children in this study begin to make this transfer.
Implications for intervention
Using captions for young at-risk readers can be an effective supplemental literacy tool in
supporting word recognition, content comprehension and code-related literacy skills after
viewing just six different educational TV programmes one time. Linebarger (2001) found
this to be true with carefully constructed scripts written with maximal reading supports
embedded. The stimuli used in this study generalise Linebarger’s (2001) findings to
commercially available, high-quality educational programming and the feasibility of
captions as an easy-to-use, pervasive and scalable intervention. Coupling our findings with
Koskinen et al.’s (1997) study of continuous caption use across many types of programming
in the home provides compelling evidence for the power of this nearly universal resource
available on every television set manufactured in the United States since July 1993.
Limitations
Several limitations must be considered when interpreting these findings. First, the small
sample size warrants caution in the generalisability of the findings beyond these
populations. However, because these children are generally at higher risk for delays in
literacy skills, the current positive findings contribute additional support for the benefits
of captions and, more generally, on-screen print to the extant literature featuring hearing-
impaired, learning-disabled and English language learners (e.g. Adler, 1985; Koskinen
et al., 1986; Linebarger, 2001; Linebarger et al., 2004; Neuman & Koskinen, 1992; Parks,
1994). In these populations, using captions and on-screen text was related to
improvements in various reading abilities as well as an increased desire to use television
as a literacy tool. Second, it is important to remember that the viewing environment of the
study (i.e. small groups in empty classrooms) may have somehow altered how the
children experienced the medium. Although all children viewed in the same environment,
it is possible that these findings would not translate in the same manner in a more
CAPTIONED TELEVISION AND LITERACY 163
rUnited Kingdom Literacy Association 2009
common viewing environment such as the child’s home. Koskinen et al.’s (1997) home
study suggests that it would.
Conclusion
Overall, these results support the developing body of evidence that early readers can learn
to read and gain a clearer understanding of new words while viewing existing children’s
educational programming with print on-screen (Linebarger, 2001; Linebarger et al., 2004;
Prince et al., 2002). Interestingly, these results also suggest that television captioning
increases attention to and subsequent comprehension of television content. If the goal is
to help children learn to recognise and subsequently understand the meaning of new
words as well as transfer specific literacy content into more generalised literacy skills,
then turning the closed captions option on while children are watching television at home
or in school is a good starting point.
Notes
1. Researchers found a curvilinear relationship between parents’ education and child screen media use (i.e.
television, videos/DVDs and movies); that is, children whose parents either had equivalent to or less than a
high school diploma or at least a college degree watched the most television while those whose parents had
some college watched the least (i.e. high school diploma or less – 4 hours 23 minutes; some college – 3 hours
46 minutes; college or greater – 4 hours 20 minutes; Roberts et al., 2005).
2. Captions for children’s programmes are not verbatim; rather, they are edited, typically for beginning or easy
readers (Media Access Group/WGBH, http://main.wgbh.org/wgbh/pages/mag/services/captioning/faq/sugg-
styles-conv-faq.html).
References
Adler, R. (1985). Using closed-captioned television in the classroom. In L.B. Gambrell & E. McLaughlin (Eds.),
New directions in reading: Research and practice. (pp. 11–18). Silver Spring, MD: Yearbook of the State of
Maryland International Reading Association.
Anderson, D.R., Huston, A.C., Schmitt, K.L., Linebarger, D.L. & Wright, J.C. (2001). Early childhood
television viewing and adolescent behavior: The recontact study. Monographs of the Society for Research in
Child Development, 66, 1–143.
Anderson, D.R. & Lorch, E.P. (1983). Looking at television: Action or reaction? In J. Bryant & D.R. Anderson
(Eds.), Children’s understanding of television: Research on attention and comprehension. (pp. 1–33). San
Diego, CA: Academic.
Anderson, D.R., Lorch, E.P., Field, D.E. & Sanders, J. (1981). The effects of TV program comprehensibility on
preschool children’s visual attention to television. Child Development, 52, 151–157.
Baker, L. & Brown, A.L. (1984). Metacognitive skills and reading. In P.D. Pearson, R. Barr, M. Kamil &
P. Mosenthal (Eds.), Handbook of reading research. (Vol. 1, pp. 353–394). New York: Longman.
Brooks-Gunn, J. & Duncan, G.J. (1997). The effects of poverty on children. The Future of Children, 7, 55–71.
Calvert, S.L., Huston, A.C. & Wright, J.C. (1987). Effects of television preplay formats on children’s attention
and story comprehension. Journal of Applied Developmental Psychology, 8, 329–342.
Campbell, T.A., Wright, J.C. & Huston, A.C. (1987). Form cues and content difficulty as determinants of
children’s cognitive processing of televised educational messages. Journal of Experimental Child Psychology,
43, 311–327.
Cohen, J. (1988). Statistical power analysis for the behavioral science. (2nd edn). Hillsdale, NJ: Lawrence
Erlbaum Associates.
Dickinson, D.K., McCabe, A., Anastasopoulos, L., Peisner-Feinberg, E. & Poe, M. (2003). The comprehensive
language approach to early literacy: The interrelationships among vocabulary, phonological sensitivity, and
print knowledge among preschool-aged children. Journal of Educational Psychology, 95, 465–481.
Dolch, E.W. (1948). Problems in reading. Champaign, IL: Garrard Press.
164 LINEBARGER, PIOTROWSKI and GREENWOOD
rUnited Kingdom Literacy Association 2009
Evans, M.A. & Saint-Aubin, J. (2005). What children are looking at during shared storybook reading: Evidence
from eye movement monitoring. Psychological Science, 16, 913–920.
Ezell, H.K., Kohler, F.W., Jarzynka, M. & Strain, P.S. (1992). Use of peer-assisted procedures to teach QAR
reading comprehension strategies to third-grade children. Education and Treatment of Children, 15, 205–227.
Federal Communication Commission (1999). Fact sheet on closed captions. Retrieved 28 January 2009, from
http://www.fcc.gov/mb/facts/ccfactsh.html
Feitelson, D. & Goldstein, Z. (1986). Patterns of book ownership and reading to young children in Israeli school-
oriented and nonschool-oriented families. Reading Teacher, 24, 924–930.
Fisch, S.M. (2000). A capacity model of children’s comprehension of educational content on television. Media
Psychology, 2, 63–91.
Fisch, S.M. (2004). Children’s learning from educational television. Mahwah, NJ: Lawrence Erlbaum.
Frey, E.B., Kress, J.E. & Fountoukidis, D. (2000). The reading teacher’s book of lists. (4th edn). Hoboken, NJ:
Jossey-Bass.
Good, R.H. & Kaminski, R.A. (2002). Dynamic indicators of basic early literacy skills (DIBELS). Eugene, OR:
Institute for the Development of Educational Achievement. Retrieved 11 February 2008, from http://
dibels.uoregon.edu/
Good, R.H., Simmons, D., Kame’enui, E., Kaminski, R.A. & Wallin, J. (2002). Summary of decision rules
for intensive, strategic, and benchmark instructional recommendations in Kindergarten through third grade
(No. 11). Eugene, OR: University of Oregon.
Griffin, R. & Dumestre, J. (1993). An initial evaluation of the use of captioned television to improve the
vocabulary and reading comprehension of navy sailors. Educational Technology, 22, 193–196.
Harris, M.M. & Smith, N.J. (1987). Literacy assessment of chapter 1 and non-chapter 1 homes. Reading
Improvement, 24, 137–142.
Hayes, A.F. & Krippendorff, K. (2007). Answering the call for a standard reliability measure for coding data.
Communication Methods and Measures, 1, 77–89.
Huston, A.C. & Wright, J.C. (1989). The forms of television and the child viewer. In G. Comstock (Ed.), Public
communication and behavior. (Vol. 2, pp. 103–158). San Diego, CA: Academic Press.
Jaccard, J. (Ed.) (1998). Interaction effects in factorial analysis of variance. Thousand Oaks, CA: Sage.
Jensema, C.J., El Sharkawy, S., Danturthi, R.S., Burch, R. & Hsu, D. (2000). Eye-movement patterns of
captioned-television viewers. American Annals of the Deaf, 145, 275–285.
Jensema, C.J., McCann, R. & Ramsey, S. (1996). Closed-captioned television presentation speed and
vocabulary. American Annals of the Deaf, 141, 284–292.
Justice, L., Skibbe, L., Canning, A. & Lankford, C. (2005). Preschoolers, print, and storybooks: An
observational study using eye-gaze analysis. Journal of Research in Reading, 28, 229–243.
Kaminski, R.A. & Good, R.H. (1996). Toward a technology for assessing basic early literacy skills. School
Psychology Review, 25, 215–227.
Kendeou, P., Lynch, J.S., van den Broek, P., Espin, C.A., White, M.J. & Kremer, K. (2005). Developing
successful readers: Building early comprehension skills through television viewing and listening. Early
Childhood Education Journal, 33, 91–98.
Koskinen, P.S., Bowen, C.T., Gambrell, L.B., Jensema, C.J. & Kane, K.W. (1997). Captioned television and
literacy development: Effects of home viewing on learning disabled students. Paper presented at the Meeting
of the American Educational Research Association, Chicago, IL.
Koskinen, P.S., Wilson, R.M., Gambrell, L.B. & Jensema, C.J. (1986). Using closed captioned television to
enhance reading skills of learning disabled students. In J.A. Niles & R.V. Lalik (Eds.), Solving problems in
literacy: Learners, teachers, and researchers. (pp. 61–65). Rochester, NY: National Reading Conference.
Kulik, C.-L.C. & Kulik, J.A. (1991). Effectiveness of computer-based instruction: An updated analysis.
Computers in Human Behavior, 7, 75–94.
Lee, J., Grigg, W. & Donahue, P. (2007). The Nation’s Report Card: Reading 2007 (No. NCES 2007–496).
Washington, DC: National Center for Education Statistics, Institute of Education Sciences, US Department of
Education.
Linebarger, D.L. (2001). Learning to read from television: The effects of using captions and narration. Journal
of Educational Psychology, 93, 288–298.
Linebarger, D.L. (2006). Teaching language and literacy on television: A ready to learn report. Philadelphia:
Annenberg School for Communication, University of Pennsylvania.
Linebarger, D.L., Kosanic, A.Z., Greenwood, C.R. & Doku, N.S. (2004). Effects of viewing the television
program Between the Lions on the emergent literacy skills of young children. Journal of Educational
Psychology, 96, 297–308.
CAPTIONED TELEVISION AND LITERACY 165
rUnited Kingdom Literacy Association 2009
Linebarger, D.L. & Piotrowski, J.T. (2006). Pinky Dinky Doo: Evaluating the educational impact and appeal of
Pinky Dinky Doo on preschool children. A final report to Sesame Workshop. Philadelphia: Annenberg School
for Communication.
Linebarger, D.L. & Piotrowski, J.T. (2009). TV as storyteller: How exposure to television narratives impacts
at-risk preschoolers’ story knowledge and narrative skills. British Journal of Developmental Psychology, 27(1),
47–69.
Linebarger, D.L. & Wainwright, D.K. (2007). American Indian children, their home media and literacy
environments, and their literacy abilities. Paper presented at the Biennial Meeting of the Society for Research
in Child Development, Boston, MA.
May, F. (1998). Reading as communication: To help children write and read. Upper Saddle River, NJ: Merrill.
Mayer, R.E. & Anderson, R.B. (1991). Animations need narrations: An experimental test of a dual-coding
hypothesis. Journal of Educational Psychology, 83, 484–490.
Mogilner, A. (1992). Children’s writer’s word book. Cincinnati, OH: Writer’s Digest Books.
Najjar, L.J. (1995). A review of the fundamental effects of multimedia information presentation on learning
(No. GIT-GVU-95-20). Atlanta, GA: Georgia Institute of Technology, Graphics, Visualization, and Usability
Center.
Nelson, D.L., Reed, V.S. & Walling, J.R. (1976). Pictorial superiority effect. Journal of Experimental
Psychology: Human Learning and Memory, 2, 523–528.
Neuman, S.B. (1995). Literacy in the television age: The myth of the TV effect. (2nd edn). Norwood, NJ: Ablex
Publishing Corporation.
Neuman, S.B. & Koskinen, P.S. (1992). Captioned television as comprehensible input: Effects of incidental
word learning from context for language minority students. Reading Research Quarterly, 27, 241–259.
Ouellette, G.P. (2006). What’s meaning got to do with it: The role of vocabulary in word reading and reading
comprehension. Journal of Educational Psychology, 98, 554–566.
Paivio, A. (1975). Coding distinctions and repetition effects in memory. In G.H. Bower (Ed.), The psychology of
learning and motivation. (Vol. 9, pp. 179–214). New York: Academic.
Paivio, A. & Csapo, K. (1973). Picture superiority in free recall: Imagery or dual coding? Cognitive Psychology,
5, 176–206.
Paivio, A., Rogers, T.B. & Smythe, P.C. (1968). Why are pictures easier to recall than words? Psychonomic
Science, 11, 137–138.
Parks, C. (1994). Closed captioned TV: A resource for ESL literacy education. Washington, DC: National
Clearinghouse for ESL Literacy Education.
Prince, D.L., Grace, C., Linebarger, D.L., Atkinson, R. & Huffman, J.D. (2002). Between the Lions Mississippi
literacy initiative: A final report to Mississippi Educational Television and WGBH Educational Foundation.
Starkville, MS: The Early Childhood Institute, Mississippi State University.
Rice, M.L., Huston, A.C. & Wright, J.C. (1982). The forms of television: Effects on children’s attention,
comprehension, and social behavior. In D. Pearl, L. Bouthilet & J.B. Lazar (Eds.), Television and behavior:
Vol. 2. Ten years of scientific progress and implications for the 80’s. (pp. 24–38). Washington, DC:
Government Printing Office.
Roberts, D.F., Foehr, U.G. & Rideout, V.R. (2005). Generation M: Media in the lives of 8–18 year-olds. Menlo
Park, CA: Henry J. Kaiser Family Foundation.
Sadoski, M. & Paivio, A. (2001). Imagery and text: A dual coding theory of reading and writing. Reading and
Writing, 16, 259–262.
Shadish, W.R., Cook, T.D. & Campbell, D.T. (2001). Experimental and quasi-experimental designs for
generalized causal inference. New York: Houghton-Mifflin.
Shea, P. (2000). Leveling the playing field: A study of captioned interactive video for second language learning.
Journal of Educational Computing Research, 22, 243–263.
Snow, C.E., Burns, M.S. & Griffin, P. (Eds.) (1998). Preventing reading difficulties in young children.
Washington, DC: National Academy Press.
Statistical Abstracts (2007). Washington, DC: Government Publishing Office.
Taylor, S.E., Frackenpohl, H., White, C.E., Nieroroda, B.W., Browning, C.L. & Birsner, E.P. (1989). EDL core
vocabularies in reading, mathematics, science, and social studies. Orlando, FL: Steck-Vaughn.
Uchikoshi, Y. (2006). Early reading development in bilingual Kindergarteners: Can educational television help?
Scientific Studies of Reading, 10, 89–120.
Vaala, S.E., Lapierre, M.A. & Linebarger, D.L. (2009). Eye movements of children and adults reading onscreen
print in educational programming. Poster presented at the Annual Convention of the Association for
Psychological Science, San Francisco, CA.
166 LINEBARGER, PIOTROWSKI and GREENWOOD
rUnited Kingdom Literacy Association 2009
Whitehouse, A.J.O., Maybery, M.T. & Durkin, K. (2006). The development of the picture-superiority effect.
British Journal of Developmental Psychology, 24, 767–773.
Wright, J.C., Huston, A.C., Murphy, K.C., St. Peters, M., Pinon, M., Scantlin, R. et al. (2001). The relations of
early television viewing to school readiness and vocabulary of children from low income families: The early
window project. Child Development, 72, 1347–1366.
Zill, N., Moore, K.S., Smith, E.W., Stief, T. & Coiro, M.J. (1995). The life circumstances and development of
children in welfare families: A profile based on national survey data. In P.L. Chase-Lansdale & J. Brooks-
Gunn (Eds.), Escape from poverty: What makes a difference for children? (pp. 38–59). New York: Cambridge
University Press.
Received 12 May 2009; revised version received 12 May 2009.
Address for correspondence: Deborah Linebarger, Annenberg School for Communica-
tion, University of Pennsylvania, 3620 Walnut Street, Philadelphia, PA 19104, USA.
E-mail: dlinebarger@asc.upenn.edu
CAPTIONED TELEVISION AND LITERACY 167
rUnited Kingdom Literacy Association 2009