Content uploaded by JOSHUA F. LAWRENCE
Author content
All content in this area was uploaded by JOSHUA F. LAWRENCE on Nov 04, 2015
Content may be subject to copyright.
Lawrence/Hinga/Mahoney/Vandell: Summer Activities, IJREE Vol. 3, Issue 1/2015, pp. 71–93
Summer Activities and Vocabulary
Development: Relationships Across Middle
Childhood and Adolescence
Joshua F. Lawrence, Briana M. Hinga, Joseph L. Mahoney, and
Deborah Lowe Vandell
Abstract: This paper examines the relation between children’s summer activities before fourth
through sixth grade and their vocabulary knowledge in fth grade and at age fteen using the NICHD
SECCYD dataset (N = 1,009). We used OLS regression and propensity score analyses to understand
how children’s summer reading, library visits, participation in enrichment classes, and unsupervised
time predicts their vocabulary knowledge. Propensity score matching and OLS analyses show that time
spent reading predicts vocabulary during the following two years, and high levels of time allocated
to reading across three or more summers in middle childhood predicts vocabulary knowledge at age
15. OLS analyses suggest a relationship between library visits and vocabulary knowledge. There is
no short-term relationship between enrichment classes and vocabulary knowledge, although our
OLS analysis demonstrated that consistent enrollment in summer enrichment classes over three years
predicted improved vocabulary. Unsupervised time predicted poor vocabulary in both the short and
long-term.
Keywords: summer, out-of-school time, vocabulary, reading, unsupervised time
1 Introduction
Student vocabulary knowledge correlates strongly with reading comprehension
measures across grade levels (Snow, Porche, Tabors, & Harris, 2007) and is a key
component of skilled adolescent reading (Kamil, 2003; Snow & Biancarosa, 2003).
Children learn new words rapidly throughout early childhood (Anglin, 1994) and
the amount and quality of home language exposure predicts children’s vocabulary
knowledge (Hart & Risley, 1995; Pan, Rowe, Singer, & Snow, 2005). As children
progress through elementary school they begin to learn more words from explicit in-
struction at school and text than from family or peer discourse. In the summer, fam-
ilies have more discretionary time to allot to preferred activities. The current paper
explores the relationship between children’s summer activity and vocabulary knowl-
edge, in both the short and the long term (at age 15). We examine the time students
(N = 1,009) spend reading, visiting the library, engaging in supervised enrichment
International Journal for Research on Extended Education, Volume 3/2015
72
activities, and unsupervised with friends in relation to their vocabulary scores in fth
grade and at age fteen using the NICHD SECCYD dataset.
Summer time. Summer is a time when many students, especially students from
low-income homes, struggle to maintain learning trajectories established during the
school year (Alexander, Entwisle, & Olson, 2001). Heyns (1978) found that sixth
and seventh graders (N = 2978) learned vocabulary at a higher average rate during
the school year than they did during the summer. She also found that out-of-school
activities and differences in family socio-economic status accounted for differenc-
es in summer vocabulary learning, but not vocabulary learning during the school
year (during which time all students were receiving instruction). Lawrence (2009)
found that sixth-grade (n = 87) and seventh-grade (n = 104) students’ vocabulary
knowledge (measure on the Group Reading and Diagnostic Evaluation; Williams,
2000) decreased during the summer. In separate longitudinal analysis of an aca-
demic vocabulary intervention, middle-school children in both treatment (n = 757)
and comparison groups (n = 204) showed marked decline in their knowledge of
high-leverage academic words during summer months (Lawrence, Capotosto, Bra-
num-Martin, White, & Snow, 2012). These ndings mirror results in related literacy
domains (Carver, 1994; Cooper, Nye, Charlton, Lindsay, & Greathouse, 1996; En-
twisle, Alexander, & Olson, 1997; Kim, 2004; Kim & White, 2008). A multi-year
study of student learning across early grades suggests learning differences during the
summer are cumulative, and that these cumulative differences explain the reading
achievement gap (Alexander, Entwisle, & Olson, 2001).
Understanding which activities are most likely to help students continue to learn
during the summer months is difcult because students who struggle during the sum-
mer are also usually the least prepared at school entry and had the least support
during the school year. Although many studies control for well-known predictors of
children’s vocabulary growth such as maternal education, family socio-economic
status, home literacy environment and school year activity in OLS regressions, sta-
tistical controls do not necessarily guard against selection bias. If high-income fam-
ilies make up most or all of the subsample that engage in enrichment activities, for
example, an OLS model might suggest a relationship between enrichment activities
and student achievement that is driven by many factors related to family wealth rath-
er than the enrichment classes per se. In this paper we use propensity score matching
as a robustness check to guard against selection bias.
This study examines how reading, library use, enrichment activities, and unsu-
pervised time predict vocabulary outcomes. We also test how cumulative summer
persistence in each activity is related to vocabulary knowledge at age 15. Each of
these activities is common during the summer, and have been explored as predictors
of vocabulary growth.
Reading. Researchers have argued that reading increasingly drives student word
learning as they get older (Nagy, Herman, & Anderson, 1985). Firstly, the density
of new words that children meet in text increases as they expand their reading diet
to include more expository texts in upper elementary and middle grades (Gardner,
2004). This means children are more likely to encounter new words in reading than
in discussion at this age. Secondly, older children are better able to infer the meaning
of new words encountered in text (Swanborn & de Glopper, 1999). Out-of-school
J. F. Lawrence, B. M. Hinga, J. L. Mahoney & D. Lowe Vandell: Summer Activities 73
reading is correlated with vocabulary knowledge. Anderson, Wilson, and Fielding
(1988) gathered self-reported daily activity logs from 155 fth graders for 26 weeks
and found reading books was associated with improvement on a vocabulary check-
list measure even after controlling for second grade reading achievement. Lawrence
(2009) found that students’ self-report of time spent reading narrative and expository
texts during the summer was related to improved vocabulary scores for better read-
ers but not for less skilled ones. Heyns (1978) found that summer reading offset the
summer setback of middle schoolers in her comprehensive analysis. Recognition
tests are an alternative measure of reading amount and correlate with vocabulary
knowledge (Cunningham & Stanovich, 1990; Cunningham & Stanovich, 1991).
Cunningham and Stanovich (1991) found that reading and receptive vocabulary cor-
related moderately (r = .46, p <.05) in a sample of fourth-, fth-, and sixth-grade
students (N = 134). Allen, Cipeilewski and Stanovich (1992) used both activity pref-
erence questionnaires and title recognition tasks and found these measures correlated
with fth-grade students’ (N = 63) vocabulary knowledge as measured by the Pea-
body Vocabulary Test and two checklist measures. For instance, students who report-
ed reading more books did better than peers on the vocabulary checklist measures
(r = .41, p <.05). In a study that recruited its participants (n = 1687) from ethnically
diverse elementary schools, Kim (2004) found that summer readers improved more
on the Stanford Achievement Test of reading than less-frequent readers. Kim and
White (2008) randomized students to three intervention conditions and found that
books plus instructional scaffolding resulted in improved literacy outcomes for stu-
dents.
There are certainly individual differences in how well students learn new words
from independent summer reading. Lawrence (2012) found that summer reading
did not offset predicted vocabulary setback for sixth and seventh grade students (N
= 278) in a longitudinal model controlling for grade level, baseline standardized
scores, gender, and home-language status. Kim and Guryan’s (2010) study of fourth-
grade students (N = 370) included measures of vocabulary knowledge, and found
that student participation in a summer reading program did not result in improved
vocabulary or comprehension scores. In a randomized trial, Kim (2006) found that
participating in a summer reading program, (which included reading instruction and
texts provided to the student during summer months) resulted in improved reading,
but was especially helpful for less-uent readers and students with fewer books at
home.
This study extends the literature about summer reading and vocabulary. For one,
we explore both short- and long-term gains associated with summer reading. Addi-
tionally, we analyze the impact of summer activities over multiple summers to test
whether the cumulative impact of activities is related to later vocabulary gains. Fur-
thermore, we use propensity score matching to compare differences between groups
of individuals who read different amount despite having the same statistical propen-
sity to read (based on key characteristics).
Library visits. According to the American Library Association (2000), 94% of
libraries surveyed throughout the US provide study space, 95% of libraries offer
summer reading programs, and 89% of libraries offer story hours – each of which
are provisions linked to academic achievement (Celano & Neumann, 2001). Kim
International Journal for Research on Extended Education, Volume 3/2015
74
(2004) found that access to libraries during the summer predicted improved reading
outcomes (controlling for baseline achievement) and there was an interaction be-
tween access and race such that Black students beneted even more than other stu-
dents from summer access to texts. Our study adds to the sparse research on library
patronship by examining the relationship between library visits and vocabulary, and
especially how regular patronship over many summers relates to adolescent vocab-
ulary knowledge.
Enrichment. Enrichment activities are of interest because they offer opportu-
nities for aural vocabulary exposure and rich discussion. Enrichment activities, in
this paper, refer to courses or programs that promote learning through recreational
means. For example, woodworking courses and hands-on science programs consti-
tute enrichment courses. In some respects, participation in these activities is similar
to school attendance. For instance, these activities are likely to provide students with
opportunities to talk and work with adults in contexts that facilitate use of special-
ized language in completing problem-solving tasks. Thus, while summer enrichment
activities may not provide rich opportunities to encounter new words in text or learn
from direct vocabulary instruction, they may provide opportunities for discussion
and new experiences. Although there is a rich research literature related to discus-
sion and reading outcomes in school contexts (Lawrence & Snow, 2010; Murphy,
Wilkinson, Soter, Hennessey, & Alexander, 2009), much less is known about how
child-adult discussion in summer or enrichment settings might support student word
learning.
Unsupervised Time with Peers. Unsupervised time may provide opportunities
for peer-to-peer discussion. During the school year, unsupervised time has been as-
sociated with mostly negative academic outcomes, however most research has been
conducted on adolescent samples. Unsupervised time has been linked to behavioral
and academic problems (e.g., Mahoney & Parente, 2009; Richardson , Radzisze-
wska, Dent, & Flay, 1993). Unsupervised time is more problematic (i.e., it leads to
delinquency) when peers are present (Osgood and Anderson, 2004; Osgood, Wilson,
O’Malley, Bachman Johnston, 1996; Warr, 2005). The current paper extends the in-
vestigation into the relationship between unsupervised time with peers by adding to
the small body of literature on the implications of unsupervised time in elementary
school.
One of the few studies involving unsupervised time in elementary school aged
children found that third grade children who spent time unsupervised actually earned
higher grades and scored higher on standardized test scores than children attend-
ing low quality after-school programs (Vandell & Corasaniti, 1988). Also, there
is evidence that peer-to-peer discussion facilitates vocabulary learning (Cekaite,
Blum-Kulka, Grøver, & Teubal, 2014). The current study adds to the small body of
literature investigating the link between unsupervised time with peers and vocabu-
lary development.
J. F. Lawrence, B. M. Hinga, J. L. Mahoney & D. Lowe Vandell: Summer Activities 75
The Present Study
The current study examines the link between time spent reading, visiting the li-
brary, taking enrichment classes, and being unsupervised during the summers before
fourth through sixth grade and vocabulary knowledge. Vocabulary is measured by
Woodcock Johnson Test Picture Vocabulary (PV) in 5th grade and at age 15. Because
activity involvement and academic achievement are dynamically related over time
(Posner & Vandell, 1999), stability and change of summer activity involvement is
measured and accounted for here. The current study includes a longitudinal explora-
tion of whether cumulative participation in each activity across summers is related to
performance on delayed vocabulary scores and uses propensity score matching as a
guard against selection bias. If child outcomes are determined by characteristics that
differ between those who select to participate in given activities versus those who do
not, propensity score matching ensures comparison between groups of students who
do not statistically differ on key observable characteristics and acts as our robustness
check. Our research questions are:
(1) Does participation in each of the above activities during summers before fourth
and/or fth grade predict PV tests scores in fth grade?
Based on current knowledge, we hypothesize that reading, library use, and enrich-
ment will predict higher vocabulary scores in fth grade. Because of inconclusive
ndings surrounding unsupervised time with peers in elementary school and vo-
cabulary, we do not have a prediction of whether unsupervised time with peers will
predict lower or higher vocabulary scores.
(2) Does participation in particular types of activities during one, two, and/or three
summers (compared to zero summers) before fourth through sixth grade sum-
mers predict PV test scores at age fteen?
We hypothesize that more reading, library use, and enrichment will predict higher
vocabulary scores at age fteen. We do not have a prediction of whether unsuper-
vised time with peers will predict lower or higher vocabulary scores. We do not
have a prediction of whether unsupervised time with peers will predict lower or
higher vocabulary scores.
(3) Do students who participate in an activity for several summers improve more
than those how don’t meet a participation thresh hold?
Because studies on cumulative activity involvement and vocabulary are lacking we
do not have specic hypotheses for the number of summers associated with vocab-
ulary outcomes.
International Journal for Research on Extended Education, Volume 3/2015
76
2 Method
2.1 Participants
Participants in the NICHD Study of Early Child Care and Youth Development were
recruited as newborns in 1991 from hospitals in or near Little Rock, AR; Irvine, CA;
Lawrence, KS; Boston, MA; Philadelphia and Pittsburgh, PA; Charlottesville, VA;
Seattle, WA; Hickory and Morganton, NC; and Madison, WI. Of the 8,986 mothers
who gave birth during the sampling period, 5,416 (60%) met eligibility requirements
and agreed to be contacted. From that pool, a conditionally random sample of 1,364
were included in the study pool which attempted to mirror the demographics of the
overall eligible sample, including: 24% ethnic minority children; 11% mothers who
had not completed high school; and 14% single family homes. Of these 1,364 chil-
dren, 1,009 remained in the study until they were 15 years old. A detailed description
of participant selection can be found in several publications (see NICHD ECCRN,
2005 for complete details) as well as on the National Institute of Child Health and
Human Development (NICHD) Study of Childcare and Youth Development website
(https://secc.rti.org).
A total of 992 children completed vocabulary tests in 5th grade, and data con-
tributed by these children are used in the rst set of analyses. For the second set of
analyses (RQ2 and 3), we use data collected from 889 children who also completed
vocabulary tests at age 15.
2.2 Measures
Summer Activity Participation. During the fourth, fth, and sixth grade school
years, mothers reported their children’s previous summer’s activity participation.
Mothers indicted the frequency that their child “read a book, magazine or newspa-
per” and “visited a library”. Response options for reading and library use ranged
from “less than once per month” to “almost every day” on a six-point scale. Parents
also reported how many weeks their child “attended an enrichment class (e.g. for-
eign language) or program for recreational learning activities such as woodworking,
hands-on science projects, art, performing arts, etc.” Lastly, parents were asked how
much time their child spent “out with friends without an adult supervising.” Re-
sponse options for enrichment and unsupervised time with peers ranged from “none”
to “8 weeks or more” along a six point scale. See Table 1 for a complete summary
of category distributions.
Activity participation responses were collapsed into two categories: high and
low activity levels. If past literature provided insight into the minimum level of each
activity which lead to improved literacy, we used criteria from existing research. If
there is no empirical base for choosing a threshold of activity participation, an at-
tempt was made to create a roughly equal distribution between groups in our data by
examining the frequency of responses.
J. F. Lawrence, B. M. Hinga, J. L. Mahoney & D. Lowe Vandell: Summer Activities 77
Summer reading groups were created by identifying students who read a few
times per week or more (i.e., the “high” group) and those who read one time per
week or less (i.e., the “low” group). This cut was made because benets of reading
occur when children engage in independent reading more than once a week (Kane,
2004). Between 64% and 68% of students were categorized as “high” readers each
summer.
Library patronage is understudied. It is not clear what threshold of library pa-
tronage is associated with improved vocabulary. We designated students who went
to the library at least 2–3 times per month as frequent library patrons (in the “high”
group). Between 32% and 41% of students were identied as active patrons each
summer.
Current research suggests relatively low levels of unsupervised time experienced
by children in the United States; even one unsupervised period a month could be
considered a high level (Mahoney & Parente, 2009). The current study categorized
students having experienced at least one period of unsupervised time a month as
frequently unsupervised and those who had not as infrequently unsupervised. The
group of “highly unsupervised” students was between 38% and 50% of the sample
each summer using this criterion.
There is sparse empirical research on the impact of enrollment in summer enrich-
ment courses. We wanted to split the distribution as evenly as possible, so we catego-
rized “high” enrichment participation as equal to any level of enrichment course ac-
tivity, and low participation as no participation. This cut off resulted in between 23%
and 26% of students being grouped into the high enrichment category each summer.
In addition to considering what levels of activity participation are used to dis-
tinguish between high and low levels of participation at each level, our analyses in-
vestigate the relationship between participation levels across multiple summers and
their vocabulary scores. Table 2 presents the percentage of students who participated
at a high level of each activity for either zero, one, two, or three summers during the
summers before fourth through sixth grade (under the heading high levels of summer
activity).
Vocabulary Measure. The Woodcock-Johnson Psycho-educational battery Test
of Achievement was used to measure children’s Picture Vocabulary (PV) scores in
fth grade and at age fteen. This vocabulary test measures verbal comprehension
(i.e., naming pictured objects). This task asks children to identify one of four pic-
tures that matches a word spoken by the examiner. Normative data for PV scores
allows for standardization and comparison of scores across time (McGrew, Werder,
& Woodcock, 1991; Woodcock, 1990). A person’s standard vocabulary score will
stay the same if their vocabulary increases at a standard rate across time. Table 2
demonstrates the mean scores of the study sample remains within a half of a standard
deviation of the normed score across all waves; vocabulary growth in this sample is
roughly similar to the norming sample.
Control variables. Because summer activity participation was not randomly as-
signed to children, the current study takes careful steps to control for confounding
variables that may be related to both activity participation and vocabulary scores.
The following three sections describe possible confounds which were controlled for
International Journal for Research on Extended Education, Volume 3/2015
78
in Ordinary Least Squares (OLS) regression equations and used as matching varia-
bles in analyses using propensity score matching.
Vocabulary. Third grade vocabulary test scores were included in each analysis to
control for vocabulary performance not long before the rst summer of interest (the
summer before fourth grade).
Child and Family Characteristics. Maternal education was reported by the
child’s mother when the child was one month old (Table 2). Average number of
years of maternal education (M = 14.23) indicates that on average mothers complet-
ed a little over two years of school after 12th grade. Child gender was reported by
the child’s mother when the child was 24 months old; 48% of the sample is female.
Ethnicity was coded as either white or non-white; 80% of the sample is white. The
family income-to-needs ratio is based on the total family income divided by the pov-
erty-level income for that family size based on federal guidelines. Scores between 0
and 1 indicate poverty, scores between 1.1 and 1.9 indicate near poverty, and scores
greater than 1.9 indicate non-poor. The mean income to needs ratio of the sample is
substantially above poverty level (M = 4.5). Finally, mothers reported the number of
parents in the home when the child was in third grade. Eighty percent of the children
in this sample lived in two parent homes.
School-year activities. The current study aims to measure how summer activity
involvement relates to vocabulary knowledge independent of school-year partici-
pation. To clarify the inuence of school-year and summer activities, third grade
school-year activities that paralleled summer activities were controlled for. To con-
trol for the inuence of school-year reading practices on test scores, the home liter-
acy score during the school year was controlled for in analyses involving summer
reading and library visits. The home literacy score was computed as the sum of
points assigned to nine items related to the child’s home literacy environment (Grif-
n & Morrison, 1997). The score was based on the mothers’ answers to nine survey
items related to the following: television watching; library card use; newspaper sub-
scription score; adult magazine subscription; child magazine subscription; mother
reads to self; adult reads to self; someone reads to child; and books owned by child.
Each of the nine items was scored from 0 to 2 points, with 2 indicating a more pos-
itive literacy environment. Total scores range from 0–17. The home literacy score
was used as a school-year control of library use.
We used the After School Time Use Child Interview, a modication of the time
use interview used by Posner and Vandell (1994, 1999), to separate the inuence of
school-year and summer-time enrichment and unsupervised time. A guided recall
format was used to obtain information about children’s weekday afternoons during
the third grade school year. For each fteen minute interval from the end of the
school day to 6:00pm, children were asked to report how they spent their time. The
interview was completed with each child up to three times in third grade. To allow
for comparisons across children, children’s time use across twenty eight recorded
activities were summed and then scaled to twelve intervals per interview to allow
comparisons to be made across children. Values for academic enrichment ranged
from 0 to 9 intervals per day. Values of unsupervised time range from 0 to 12 inter-
vals per day.
J. F. Lawrence, B. M. Hinga, J. L. Mahoney & D. Lowe Vandell: Summer Activities 79
School-year and summer activities (i.e., the independent variables of interest)
are only modestly correlated with each other (Table 3).
2.3 Data Analysis
Analyses corresponding to the rst research question (RQ1) illuminate relations be-
tween summer activity participation in fourth or fth grade and fth grade vocabu-
lary scores. Analyses corresponding to the second (RQ2) and third (RQ3) research
questions illuminate relations between participation across summers between fourth
through sixth grade and vocabulary scores at age 15.
RQ1. Summer activity participation predicting vocabulary tests in grade
ve. The rst set of analyses tested hypothesized associations between participa-
tion in specied activities during the summers before fourth and fth grade and
tests of vocabulary in fth grade. These analyses were conducted in two steps. First,
fth-grade vocabulary tests scores of children who participated in each activity (i.e.,
reading, library visits, enrichment courses, and unsupervised time with peers) during
fourth and/or fth grade summers were compared with scores of students who did
not participate in each summer activity. Because participants were not randomly
assigned to activity participation, control variables included: third grade PV scores;
gender; ethnicity; maternal education; single parent status; and family income to
needs ratio. Additionally, independent school-year activity participation levels were
also included as control variables (as described above).
As a robustness check against selection bias, propensity score matching was
performed to match individuals who participated in each summer activity during
the summers before fourth and fth grade to those who did not participate in the
activity but had a similar probability of participation. Using PSMATCH2 (Leuven &
Sianesi, 2003) to perform 1-to-1 propensity score matching with replacement, pro-
pensity scores were developed to predict participation in each summer activity using
control variables (i.e., third grade PV scores; gender; ethnicity; maternal education;
single parent status; family income to needs ratio and participation in each activity
during the school year). To determine whether each summer activity predicted tests
of vocabulary, children who participated in each activity during each summer were
compared to propensity-matched individuals not involved in the activity of interest
during that summer.
RQ2. Cumulative summer activity participation predicting vocabulary
scores at age fteen. The second set of analyses tested the hypothesized associa-
tions between vocabulary scores at age fteen between children who participated
in each activity during either one, two, or three summers (during fourth through
sixth grade) versus those who do not participate in the activity during any of these
summers. Again, because participants were not randomly assigned to activity partic-
ipation, a list of confounds were controlled for in this initial regression analysis (see
list of control variables above). Propensity score matching was not performed in this
case because the variables of interest (number of summers at high levels of activity)
were not dichotomous.
International Journal for Research on Extended Education, Volume 3/2015
80
RQ3. Threshold analysis of summer activity participation predicting vocab-
ulary scores at age fteen. The third set of analyses tested whether a minimum
number of summers (i.e., a threshold) signicantly related to test scores when com-
paring children who participated in activities above and below this threshold. The
threshold of activity participation across summers was determined as the least num-
ber of summers associated with signicantly different test scores for children in each
activity compared to those who did not participate in the activity at all in the analyses
for RQ2 above. This threshold was used to determine two groups of children for each
activity (i.e., those who participated in the activity at or above the threshold versus
those who did not).
As a robustness check, propensity score matching was performed to match indi-
viduals who participated in each summer activity at or above the threshold to those
who did not but had a similar probability of activity participation up to the threshold.
Using PSMATCH2 (Leuven & Sianesi, 2003) to perform 1-to-1 propensity score
matching with replacement, propensity scores were developed predicting participa-
tion in each summer activity of interest using the control variables specied above
as matching variables. To determine whether each “threshold” of summer activities
predicted WJ-R tests of vocabulary test scores, children at or above the threshold
were compared to propensity-matched individuals not involved in the activity of
interest up to the threshold level.
3 Results
3.1 RQ1. Summer Activity Participation Predicting Vocabulary Scores
in 5th Grade
Table 4 provides results from OLS and propensity score matching analysis.1
Reading. In both the OLS and propensity score matching (PSM) analyses, chil-
dren in the high reading exposure groups during fourth or fth grade summers scored
signicantly higher on vocabulary tests in fth grade (OLS: b = 3.04, p < .01; b =
2.68, p < .01.; PSM: b = 3.21, p < .05; b = 2.95, p < .001).
Library visits. There was a positive signicant relationship between library use
during the summer of fourth grade and vocabulary in fth grade in the OLS regres-
sion analyses (b = 1.84, p < .05; b = 1.90, p < .05.), but this relation was not evident
in the propensity score analysis.
Enrichment. There were no signicant differences on fth grade vocabulary
tests between children who attended summer enrichment classes during summers
before fourth or fth grade and those who did not.
1 Children participating in each summer activity were well matched to non-participants after propensity score
matching. With few exceptions, bias in matching variables described above was reduced after propensity
score matching between children involved in each summer activity versus those not involved in each activity.
Comparisons after matching are described in the following section. Signicance tests for an interaction
between participation in each activity and maternal education were not found to be signicant for any of the
below analyses and were therefore omitted.
J. F. Lawrence, B. M. Hinga, J. L. Mahoney & D. Lowe Vandell: Summer Activities 81
Unsupervised Time with Peers. Children who were unsupervised with peers
during the summer before fth grade displayed signicantly lower vocabulary tests
in fth grade in both the OLS regressions (b = -1.87, p < .05) and the PSM analysis
(b = -2.25, p < .05). There were no signicant differences in the vocabulary tests of
children in the high and low unsupervised groups during the fourth grade summer.
3.2 RQ2. Cumulative Summer Activity Participation Predicting Vocabulary
Scores at Age Fifteen
Next, we describe associations between the number of summers that children partic-
ipated in each activity between fourth through sixth grade summers and their vocab-
ulary test scores at age 15. These results tell us about relationships between summer
activities and vocabulary development for students in our sample, and help us es-
tablish a threshold for cross-summer activity levels we use in RQ3. Unstandardized
coefcients and effects sizes (calculated by dividing the coefcient by the grand
vocabulary standard deviation [SD = 14.8]) are reported on Table 5.
Reading. Vocabulary tests were estimated for children who had high reading
participation for one, two, or three summers (versus zero summers) between fourth
through sixth grade. Children who read regularly during at least three summers
scored signicantly higher on vocabulary tests at age 15 years (d = 0.41, p > .001)
than children who did not regularly read at high levels between fourth through sixth
grade.
Library visits. Children who regularly visited the library across all three sum-
mers scored higher on vocabulary test scores at age 15 than those who did not regu-
larly visit the library during any summer (d = 0.22, p < .01).
Enrichment. Children who participated in enrichment courses for three sum-
mers demonstrated better vocabulary knowledge at age 15 (d = .49, p < .01) than
those who did not attend enrichment courses regularly during any summer.
Unsupervised Time. Children who were unsupervised for two or three summers
between fourth through sixth grade scored lower on vocabulary tests at age fteen
(d = -0.18, p < .05 and d = -.37, p < .001 respectively) than children who were not
regularly unsupervised during any summer between fourth through sixth grade.
3.3 RQ3. Threshold Analysis of Cumulative Summer Activity Participation
Predicting Vocabulary Scores at Age Fifteen
RQ3 results illuminate differences between individuals who participated in each
activity above and below the threshold number of summers associated with sig-
nicantly different test scores, as determined in response to RQ2. Unstandardized
coefcients and effects sizes for both OLS and propensity score matching analysis
are provided on Table 6.
Reading. In our last set of analysis (RQ2) we saw that students who reported
high levels of reading for three summers had better age-15 vocabulary scores than
students who did not read at high levels during any summer. Therefore, in this set
International Journal for Research on Extended Education, Volume 3/2015
82
of analysis (RQ3) we compare students who reported reading at high levels during
three summers with those who read at high levels for only two summers or less using
OLS and propensity score matching. Both approaches indicate that those who read
during at least three summers scored signicantly higher on vocabulary tests at age
15 (OLS d = .27, p < .001; PSM d = .28, p < .001).
Library visits. Children who regularly visited the library across all three sum-
mers scored higher on vocabulary tests at age 15 than those who did not regularly
visit the library during the summer. Accordingly, three summers was determined as
the threshold for RQ3 analysis. However, although the parameter associated with
3 summers of high levels of library patronage was signicant in the OLS analysis
(OLS d = .26, p < .001), it was not in the propensity score matching model.
Enrichment. Three summers was used as the threshold for analyzing the re-
lationship between enrichment attendance and vocabulary. OLS threshold analysis
(OLS d = .46, p < .001) suggests the importance of consistent attendance in enrich-
ment classes. However, after matching, no signicant differences in age fteen vo-
cabulary scores were found between children who did or did not regularly attended
enrichment courses for three summers.
Unsupervised Time. In both OLS and propensity score matching models, chil-
dren who were unsupervised for two summers or more scored signicantly lower on
vocabulary tests (OLS d = -0.25 , p < .001; PSM d = -0.19, p < .01) than children
who experienced zero or one summer with an unsupervised period.
4 Discussion
Findings from this study ll gaps in the literature concerning relations between sum-
mertime activity involvement and vocabulary knowledge. Specically, three issues
were addressed: (1) whether participation in specic activities during the summers
before fourth and fth grade is related vocabulary test scores in fth grade; (2) wheth-
er cumulative activity participation across summers between fourth and sixth grade
predicts vocabulary test scores at age fteen; and (3) whether there is a threshold
number of summers in each activity associated with vocabulary test scores at age 15.
The most consistent nding is that reading is an important predictor of vocab-
ulary knowledge in both the short and long term. Findings indicated that summer
reading during fourth and/or fth grade is positively associated with higher vocabu-
lary scores whereas unsupervised summer time during fth grade is related to lower
PV scores. Three summers of reading between summers before fourth through sixth
grade predict higher vocabulary scores at age 15. These ndings are consistent with
previous research showing that reading during the summer is associated with subse-
quent positive academic achievement (e.g., Kim, 2004). Current ndings add to the
literature by revealing that reading during the summer is not only associated with
short term academic achievement in grade 5, but is associated with longer term aca-
demic achievement in later adolescence as well. Further, the current paper provides
evidence that consistent reading across each summer between fourth through sixth
is the activity threshold associated with higher vocabulary at age 15. If this nding
J. F. Lawrence, B. M. Hinga, J. L. Mahoney & D. Lowe Vandell: Summer Activities 83
holds across future studies, careful attention should be paid to fostering reading op-
portunities for children across multiple summers.
The nding that unsupervised time with peers during the summer is associated
with lower vocabulary scores adds to the literature on unsupervised time during
elementary school and during the summers. Unsupervised time has previously been
associated with misconduct and problem behaviors. The current ndings suggest
that at least two summers of unsupervised time are also related to lower vocabulary
scores in later adolescence. The lack of research on the relation between unsuper-
vised time during elementary grades and academic achievement may be explained
by the fact that at least two summers of unsupervised time are required before signif-
icant differences in test scores appear.
The nding that library use and enrichment participation were only associated
with signicant ndings before propensity score matching may suggest that selection
differences were controlled through propensity score matching. The magnitude of
the effect sizes yielded by summer enrichment activity participation is noteworthy.
A review of out-of-school programs indicated that program effect sizes are strongly
related to high levels of program implementation and consistency of implementation
(Durlak & DuPre, 2008). Specically, programs that were consistently implemented
have yielded effect sizes as high as .50. Because the current study analyzed activity
dichotomously, as either at or above previously determined “effective” levels of ac-
tivity involvement (based on past literature) it makes sense that effect sizes would
be similar to effect sizes for “high levels of implementation.” The fact that these
activities are not signicant in the propensity score matching analyses suggest that
these classes may be enrolled in by families with other advantages and so selection
bias needs to be carefully considered when estimating the impact of these programs.
Limitations and Future Directions
While this study adds to the literature on summer and vocabulary development, sev-
eral limitations should be noted. First, while the sample was ethnically and eco-
nomically diverse, the NICHD dataset does not include language minority children
because the initial sample was created from a pool of English speaking mothers. A
nationally and linguistically representative sample is needed to make broader gener-
alizations about ndings.
A second limitation is that this study was not a randomly assigned experiment.
However, use of controls and propensity score matching provided a strong test of
quasi-experimental research. Importantly, school-year activities that closely matched
summer activities of interest were included as control variables in the regular regres-
sion analyses and matching variables in the propensity score matching analyses.
Inclusion of school-year activities allowed for a more precise measure of the asso-
ciation between activity involvements during the summer without confounding par-
ticipation in the activity during the school year. Future studies can similarly benet
from inclusion of school-year activities as controls if the goal is to isolate effects of
summer activities, independent of school year activity participation.
International Journal for Research on Extended Education, Volume 3/2015
84
Future research should measure activity quality. This study did not take into
account activity quality because this measure was not available. This is a limitation
because quality of out-of-school time experiences has been shown to be signicant-
ly linked to outcomes (Pierce, Hamm, & Vandell, 1999; Posner & Vandell, 1994;
Rosenthal & Vandell, 1996; Vandell et al., 2006). Similarly, the current study was
limited by the inability to account for activity content. For example, the relation
between library use and test scores may depend on the content of what children do
while at the library. As with any activity, the content of the experience (e.g., whether
a student completed research in the library or talked to friends) should be considered
in studies of summer experiences. For example, the null ndings associated with
enrichment activities could be related the fact that enrichment activities in this study
were measured as a hodgepodge of different sorts activities without knowledge of
organization level or quality. Future studies would benet from looking at whether
specic enrichment activities are related to vocabulary scores. Future studies that use
more specic measures of summer activities and specic learning processes within
the activities would be informative.
Despite limitations, this study begins to ll gaps in the literature concerning
the relation between summer experiences (during middle childhood in the summers
between fourth through sixth grade) and measures of vocabulary up to age fteen.
The current study’s ndings that specic elementary summer experiences predict
vocabulary at age fteen provide impetus for further research into understanding
potentially effective summer learning opportunities. This study indicates that stud-
ying activity involvement cumulatively across summers and over time is important
to understanding relationships between activities and outcomes over time. Overall,
current ndings t with past literature demonstrating that summer learning during
elementary school explains academic achievement into high school (Alexander et
al., 2001) and provide more information about what activities are related to a specic
measure of vocabulary development.
References
Alexander, K., Entwisle, D., & Olson, L. (2001). Schools, achievement, and inequal-
ity: A seasonal perspective. Educational Evaluation & Policy Analysis, 23(2),
171‒191.
Allen, L., Cipielewski, J., & Stanovich, K. (1992). Multiple indicators of children’s
reading habits and attitudes: Construct validity and cognitive correlates. Journal
of Educational Psychology, 84(4), 489‒503.
Anderson, R., Wilson, P. T., & Fielding, L. G. (1988). Growth in reading and how
children spend their time outside of school. Reading Research Quarterly, 23,
285‒303.
Anglin, J. (1994). Vocabulary development: A morphological analysis. Monographs
of the Society for Research in Child Development, 58(10).
J. F. Lawrence, B. M. Hinga, J. L. Mahoney & D. Lowe Vandell: Summer Activities 85
Carver, R. (1994). Percentage of unknown vocabulary words in text as a function of
the relative difculty of the text: Implications for instruction. Journal of Reading
Behavior, 26(4), 413‒437.
Cekaite, A., Blum-Kulka, S., Grøver, V., & Teubal, E. (Eds.) (2014). Children‘s peer
discourse. Peer interactions and pragmatic development in rst and second lan-
guage. Cambridge: Cambridge University Press.
Celano, D., & Neumann, S. (2001). The role of public libraries in children’s literacy
development. Pennsylvania, PA: Pennsylvania Library Association.
Cooper, H., Nye, B., Charlton, K., Lindsay, J., & Greathouse, S. (1996). The effects
of summer vacation on achievement test scores: A narrative and meta-analytic
review. Review of Educational Research, 66(3), 227‒268.
Cunningham, A., & Stanovich, K. (1990). Assessing print exposure and orthograph-
ic processing skill in children: A quick measure of reading experience. Journal
of Educational Psychology, 82(4), 733‒740.
Cunningham, A., & Stanovich, K. (1991). Tracking the unique effects of print expo-
sure in children: Associations with vocabulary, general knowledge, and spelling.
Journal of Educational Psychology, 83(2), 264‒274.
Durlak, J. A., & DuPre, E. P. (2008). Implementation matters: A review of research
on the inuence of implementation on program outcomes and the factors af-
fecting implementation. American Journal of Community Psychology, 41(3),
327‒350.
Entwisle, D., Alexander, K., & Olson, L. (1997). Children, schools and inequality.
Boulder, CO: Westview Press.
Gardner, D. (2004). Vocabulary input through extensive reading: A comparison of
words found in children’s narrative and expository reading materials. Applied
Linguistics, 25(1), 1‒37.
Grifn, E. A., & Morrison, F. J. (1997). The Unique Contribution of Home Literacy
Environment to Differences in Early Literacy Skills 1. Early Child Development
and Care, 127(1), 233‒243.
Hart, B., & Risley, T. (1995). Meaningful differences in the everyday experience of
young American children. Baltimore, MD: Brookes Publishing Company.
Heyns, B. (1978). Summer learning and the effects of schooling. New York: Aca-
demic Press.
Kamil, M. L. (2003). Adolescents and literacy: Reading for the 21st century. Wash-
ington, DC: Alliance for Excellent Education.
Kane, T. J. (2004). The impact of after-school programs: Interpreting the results of
four recent evaluations. New York: William T. Grant Foundation.
Kim, J. (2004). Summer reading and the ethnic achievement gap. Journal of Educa-
tion and Students Placed at Risk, 9(2), 169‒188.
Kim, J. (2006). Effects of a voluntary summer reading intervention on reading
achievement: Results from a randomized eld trial. Educational Evaluation and
Policy Analysis, 28(4), 335.
Kim, J. S., & Guryan, J. (2010). The efcacy of a voluntary summer book reading
intervention for low-income Latino children from language minority families.
Journal of Educational Psychology, 102(1), 20.
International Journal for Research on Extended Education, Volume 3/2015
86
Kim, J., & White, T. (2008). Scaffolding voluntary summer reading for children in
grades 3 to 5: An experimental study. Scientic Studies of Reading, 12(1), 1‒23.
Lawrence, J. (2009). Summer reading: Predicting adolescent word learning from
aptitude, time spent reading, and text type. Reading Psychology, 30(5), 445‒465.
doi: 10.1080/02702710802412008
Lawrence, J. (2012). English vocabulary learning trajectories of students whose par-
ents speak a language other than English: Steep learning and deep summer set-
back. Reading and Writing: An Interdisciplinary Journal. doi: 10.1007/s11145-
011-9305-z.
Lawrence, J., Capotosto, L., Branum-Martin, L., White, C., & Snow, C. (2012). Lan-
guage prociency, home-language status, and English vocabulary development:
A longitudinal follow-up of the Word Generation program. Bilingualism: Lan-
guage and Cognition. doi: 10.1017/S136672891100039
Lawrence, J., & Snow, C. (2010). Oral discourse and reading comprehension. In M.
Kamil, D. Pearson, E. Moje, P. Aerback & P. Mosenthal (Eds.), Handbook of
Reading Research. (Vol. IV, pp. 320‒337). London: Routledge.
Leuven, E., & Sianesi, B. (2003). PSMATCH2: Stata module to perform full Ma-
halanobis and Propensity Score Matching, version 1.2.1.
Mahoney, J. L., & Parente, M. E. (2009). Should We Care About Adolescents Who
Care for Themselves? What We Have Learned and What We Need to Know
About Youth in Self‐Care. Child Development Perspectives, 3(3), 189‒195.
McGrew, K. S., Werder, J. K., & Woodcock, R. (1991). WJ-R technical manual.
Allen, TX: DLM.
Murphy, P., Wilkinson, I., Soter, A., Hennessey, M., & Alexander, J. (2009). Exam-
ining the effects of classroom discussion on students’ comprehension of text: A
meta-analysis. Journal of educational psychology, 101(3), 740.
Nagy, W., Herman, P., & Anderson, R. C. (1985). Learning words from context.
Reading Research Quarterly, 20(2), 233‒253.
Pan, B., Rowe, M., Singer, J., & Snow, C. (2005). Maternal correlates of growth
in toddler vocabulary production in low-income families. Child Development,
76(4), 763‒782.
Pierce, K. M., Hamm, J. V., & Vandell, D. L. (1999). Experiences in After‐School
Programs and Children’s Adjustment in First‐Grade Classrooms. Child develop-
ment, 70(3), 756‒767.
Posner, J. K., & Vandell, D. L. (1994). Low‐Income Children’s After‐School Care:
Are There Benecial Effects of After‐School Programs? Child development,
65(2), 440‒456.
Posner, J. K., & Vandell, D. L. (1999). After-school activities and the development
of low-income urban children: A longitudinal study. Developmental Psychology,
35(3), 868.
Richardson, J. L., Radziszewska, B., Dent, C. W., & Flay, B. R. (1993). Relation-
ship between after-school care of adolescents and substance use, risk taking,
depressed mood, and academic achievement. Pediatrics, 92(1), 32.
Rosenthal, R., & Vandell, D. L. (1996). Quality of care at school‐aged child‐care
programs: Regulatable features, observed experiences, child perspectives, and
parent perspectives. Child development, 67(5), 2434‒2445.
J. F. Lawrence, B. M. Hinga, J. L. Mahoney & D. Lowe Vandell: Summer Activities 87
Snow, C., & Biancarosa, G. (2003). Adolescent literacy and the achievement gap:
What do we know and where do we go from here? New York: Carnegie Corpo-
ration of New York.
Snow, C., Porche, M. V., Tabors, P., & Harris, S. (2007). Is literacy enough?: Path-
ways to academic success for adolescents. Baltimore, MD: Paul H. Brookes
Publishing Co.
Swanborn, M., & de Glopper, K. (1999). Incidental word learning while reading: A
meta-analysis. Review of Educational Research, 69(3), 261‒285.
Vandell, D. L., & Corasaniti, M. A. (1988). The relation between third graders’ af-
ter-school care and social, academic, and emotional functioning. Child develop-
ment, 868‒875.
Vandell, D. L., Reisner, E. R., Pierce, K. M., Brown, B. B., Lee, D., Bolt, D., & Pech-
man, E. M. (2006). The study of promising after-school programs: Examination
of longer term outcomes after two years of program experiences. Madison, WI:
Wisconsin Center for Education Research.
Williams, K. T. (2000). Group Reading Assessment and Diagnostic Evaluation. Cir-
cle Pines, MN: American Guidance Service.
Woodcock, R. W. (1990). Theoretical foundations of the WJ-R measures of cogni-
tive ability. Journal of Psychoeducational Assessment, 8(3), 231‒258.
International Journal for Research on Extended Education, Volume 3/2015
88
Table 1. Distribution of scores along the six point scales of activity involvement
>1x/
month
≈ 1x/
month
2-3x/
month
≈ 1x/
week
few x/
week ≈ daily N Total % in
“high” group
Reading
Grade 4 2% 6% 9% 15% 29% 39% 957 68%
Grade 5 3% 7% 12% 14% 28% 36% 987 64%
Grade 6 4% 6% 11% 14% 27% 38% 949 65%
Library Visits
Grade 4 37% 23% 20% 16% 4% 1% 951 41%
Grade 5 38% 25% 16% 16% 4% 1% 975 37%
Grade 6 43% 24% 15% 13% 4% 1% 944 32%
Unsupervised Time
Grade 4 61% 4% 5% 7% 13% 9% 948 38%
Grade 5 56% 7% 7% 8% 14% 9% 982 45%
Grade 6 50% 7% 8% 11% 14% 10% 948 50%
None 1 day–
1 week 2–3 weeks 4 weeks 5–7
weeks
≥ 8
weeks NTotal % in
“high” group
Enrichment
Grade 4 74% 9% 8% 3% 3% 2% 940 26%
Grade 5 74% 10% 9% 3% 2% 2% 969 26%
Grade 6 77% 10% 7% 2% 3% 2% 937 23%
Note: The frequencies for each activity align with the frequencies on the parent questionnaire soliciting
amount of time their child spent in each activity during the previous summer. Parents were asked to
circle a number (1-6) to indicate which category of time their child spent on each activity. For “read-
ing,” “library visits,’ and “unsupervised time” category options range from “less than once a month”
to “almost every day” as indicated in the top columns. For “enrichment” category options ranged from
“none” to “8 weeks or more” as indicated by the column headers directly above “enrichment.” The
numbers in italics represent categories that are part of the “high” activity involvement group for cer-
tain analyses, whereas percentages that are not in italics represent activities part of the “low” activity
involvement for the given activity.
J. F. Lawrence, B. M. Hinga, J. L. Mahoney & D. Lowe Vandell: Summer Activities 89
Table 2. Demographic Characteristics of Children and Families
N % Mean SD Range
Demographics
Maternal education 14.23 2.51 7 to 21
Female children 1364 48%
Ethnicity: white 1364 80%
Family income/needs 985 4.5 3.88 .07 to 32
Two parent homes 1045 80% 0 to 1
Home Literacy Score 1016 10.31 3.57 1 to 17
Enrichment 1022 .45 .90 0 to 9
Unsupervised Time 1022 1.78 2.37 0 to 12
High levels of summer activity
Reading 0 Summers 889 12%
1 Summer 889 20% 0 to 1
2 Summers 889 20% 0 to 1
3 Summers 889 49% 0 to 1
Library 0 Summers 889 20%
1 Summer 889 17% 0 to 1
2 Summers 889 24% 0 to 1
3 Summers 889 39% 0 to 1
Enrichment 0 Summers 889 51%
1 Summer 889 27% 0 to 1
2 Summers 889 14% 0 to 1
3 Summers 889 7% 0 to 1
Unsupervised 0 Summers 889 33%
1 Summer 889 25% 0 to 1
2 Summers 889 20% 0 to 1
3 Summers 889 22% 0 to 1
Vocabulary
Grade 3 1014 105.47 14.8 34 to 152
Grade 5 992 103.1 14.8 29 to 155
Age 15 889 99.93 14.8 34 to158
Note: Maternal education is measured by years of schools starting in rst grade. The mean of 14.23
represents completion of 2.23 years of school after 12th grade. For analysis purposes, ethnicity was
collapsed into white versus all other ethnicities. A family income/needs ratio of 0‒1 indicates poverty,
1.1‒1.9 indicates near poverty, and greater than 1.9 indicates non-poor. Two parent homes is a measure
of how many children live with two parents (instead of only one).
International Journal for Research on Extended Education, Volume 3/2015
90
Table 3. Correlations between activity involvement during summer (left column)
and school year (top row) activities
Summer School-year
Activity Literacy Score Enrichment Unsupervised
Reading
Grade 4 .30*** .04 -.06
Grade 5 .29*** .09** -.05
Grade 6 .31*** .11*** -.06
Library
Grade 4 .22*** .10** -.08*
Grade 5 .18*** .07* -.04
Grade 6 .14*** .03 -.06
Enrichment
Grade 4 .17*** .07* .05
Grade 5 .14*** .07* -.04
Grade 6 .21*** .09** .01
Unsupervised
Grade 4 -.16*** -.06 .12***
Grade 5 -.21*** -.03 .10**
Grade 6 -.15*** -.12** .12***
Note: The numbers in italics represent matched school and summer year variables.
*p < .05; ** p < .01; *** p < .001
J. F. Lawrence, B. M. Hinga, J. L. Mahoney & D. Lowe Vandell: Summer Activities 91
Table 4. Regression Coefcients (and Standard Errors) of Summer Activity
Involvement and Vocabulary Scores in Grade 5 for OLS and
Propensity-Score Matching Analysis
OLS Regressions Propensity Score
Matching
Reading
Grade 4 3.04** (.96) 3.21* (.79)
Grade 5 2.68** (.92) 2.95*** (.82)
Library
Grade 4 1.84* (.90) -.71 (1.15)
Grade 5 1.90* (.90) .87 (1.24)
Enrichment
Grade 4 1.26 (1.02) .37 (1.51)
Grade 5 1.78 (.99) 2.10 (1.49)
Unsupervised
Grade 4 -1.32 (.93) -.12 (1.12)
Grade 5 -1.87* (93) -2.25* (1.05)
Note. Study members were categorized as involved in each activity at least once per week or not. Con-
trols variables include: 3rd grade vocabulary score; sex; mother’s education; family income to needs
ratio; whether or not the family is a singly family household; child’s race; data collection site; involve-
ment in specic activity during school year. The same variables used as controls were used as matching
variables in the propensity score matching analysis.
*p < .05.; ** p < .01; *** p < .001
International Journal for Research on Extended Education, Volume 3/2015
92
Table 5. OLS Regression Coefcients (and Standard Errors) of Summer Activity
Involvement during One, Two, or Three Summers compared to
Zero Summers (During 4th‒6th Grade)
Vocabulary Scores
OLS Coecient Eect Size
Reading
1 summer .80 (1.74) 0.05
2 summers 2.85 (1.75) 0.19
3 summers ◊ 6.02*** (1.51) 0.41
Library
1 summer -.48 (1.19) -0.03
2 summers -.26 (1.33) -0.02
3 summers ◊ 3.31* (1.36) 0.22
Enrichment
1 summer .78 (1.15) 0.05
2 summers 1.01 (1.39) 0.07
3 summers ◊ 7.26** (1.86) 0.49
Unsupervised
1 summer -1.80 (1.25) -0.12
2 summers ◊ -2.66* (1.37) -0.18
3 summers -5.44*** (1.35) -0.37
Note. Controls variables include: either 3rd grade vocabulary score; sex; mother’s education; family in-
come to needs ratio; whether or not the family is a singly family household; child’s race; data collection
site; involvement in specic activity during school year. The same variables used as controls were used
as matching variables in the propensity score matching.
◊ indicates the “threshold” or least number of summers associated with either signicantly higher or
lower test scores than children who did not participate in the activity during any summer measured.
*p < .05.; ** p < .01; *** p < .001
J. F. Lawrence, B. M. Hinga, J. L. Mahoney & D. Lowe Vandell: Summer Activities 93
Table 6. OLS and Propensity Score Matching Analysis of Activity Thresholds
Predicting Vocabulary Scores at Age 15
OLS Coecients Eect
Size Propensity Score
Matching
Eect
Size
High Reading 3.93*** 0.27 4.07*** 0.28
(3 summers vs. 0, 1, 2 summers) (1.14) (1.29)
High Library Patronage 3.92** 0.26 2.97 0.20
(3 summers vs. 0, 1, 2 summers) (1.38) (2.07)
Attending enrichment 6.76*** 0.46 4.98 0.34
(3 summers vs. 0, 1, 2 summers) (1.78) (3.35)
Unsupervised time -3.67*** -0.25 -2.75** -0.19
(3 & 4 summers vs. 0, 1 & 3 summers) (.98) (1.09)
Note. Controls variables include: either 3rd grade vocabulary score; sex; mother’s education; family in-
come to needs ratio; whether or not the family is a singly family household; child’s race; data collection
site; involvement in specic activity during school year. The same variables used as controls were used
as matching variables in the propensity score matching.
*p < .05.; ** p < .01; *** p < .001