Professional Documents
Culture Documents
Improving The General Language Skills of Second Language Learners in Kindergarten A Randomized Controlled Trial
Improving The General Language Skills of Second Language Learners in Kindergarten A Randomized Controlled Trial
To cite this article: Kristin Rogde, Monica Melby-Lervåg & Arne Lervåg (2016) Improving
the General Language Skills of Second-Language Learners in Kindergarten: A Randomized
Controlled Trial, Journal of Research on Educational Effectiveness, 9:sup1, 150-170, DOI:
10.1080/19345747.2016.1171935
Accepted author version posted online: 08 Submit your article to this journal
Apr 2016.
Published online: 20 Jun 2016.
ABSTRACT KEYWORDS
Second-language learners display poorer general language skills in the second-language learners
language used at school than their monolingual peers, which is a general language skills
concern because general language skills (vocabulary, grammar, vocabulary
language expression, and comprehension) provide the foundation for intervention
randomized controlled trial
later academic success. In a randomized controlled trial, we examined
the efficacy of an intervention to improve second-language learners’
general language skills in their second language. One hundred and
fifteen children (mean age D 5.5 years) were randomly assigned to a
business-as-usual control group or to an 18-week intervention
program conducted by kindergarten teachers during the last semester
of kindergarten. The children were assessed at pre-intervention, post-
intervention and a seven-month follow-up. The children in the
treatment group demonstrated significant improvements on a custom
measure of taught vocabulary. The language intervention program
also produced effects that generalized to standardized measures
reflecting expressive language skills. These findings suggest that
intervention programs designed to enhance second-language learners’
expressive language in the second language can be successfully
implemented in kindergarten settings before school entry.
Worrisome low academic achievement and high dropout rates among second-language
learners have been reported during the past few years (Child Trends Databank, 2011;
Organisation for Economic Co-operation and Development [OECD], 2010). The number of
second-language-learner children whose first language differs from the predominant lan-
guage of instruction where they reside has increased sharply in the United States, Europe,
and North America in recent decades (National Center for Education Statistics [NCES],
2013; OECD, 2010). A number of studies have shown that these children display poorer gen-
eral language (expressive and receptive) skills in the second language than their monolingual
peers (see Melby-Lervag & Lervag, 2014 for a review). This gap in language proficiency
between first- and second-language learners appears to persist over time (Droop & Verho-
even, 2003; Kieffer & Vukovic, 2013; Lervag & Aukrust, 2010; Lonigan, Farver, Nakamoto,
& Eppe, 2013) and cannot be solely explained by differences in socioeconomic status
CONTACT Kristin Rogde kristin.rogde@iped.uio.no Department of Education, University of Oslo, Postboks 1092 B,
Oslo, Norway.
a
University of Oslo, Oslo, Norway
Supplemental material data for this article can be accessed at http://dx.doi.org/10.1080/19345747.2016.1171935.
© Kristin Rogde, Monica Melby-Lervag, and Arne Lervag.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://
creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided
the original work is properly cited, and is not altered, transformed, or built upon in any way.
(Melby-Lervag & Lervag, 2014). Therefore, being a second-language learner itself appears to
be a risk factor for having poor general language skills.
General language skills can be understood as the linguistic forms of language that are
associated with the ability to understand and express language (i.e., vocabulary and gram-
mar), often assessed by tests of vocabulary or listening comprehension (Bornstein, Hahn,
Putnick, & Suwalsky, 2014; Foorman, Herrera, Petscher, Mitchell, & Truckenmiller, 2015;
Klem et al., 2014, Melby-Lervag & Lervag, 2014). General language skills are known to
underpin the development of reading skills, which again, are necessary for content-area
learning in all subjects and thus are an influential factor for academic success. These skills
are described as a well-known precursor to reading success that develops long before formal
reading instruction begins (Snow, Burns, & Griffin, 1998; Whitehurst & Lonigan, 1998). In
older children, general language skills and reading comprehension are highly related as well
(e.g., Foorman, Koon, Petscher, Mitchell, & Truckenmiller, 2015).
General language skills are also indispensable for successful communication with others.
Understanding others and expressing one’s thoughts and ideas through a shared language
are crucial for children’s social well-being and emotional development (Beitchman et al.,
2001; Cohen, 2001). As the number of children entering school with the language of instruc-
tion as a second language continues to grow, early intervention to prevent undesirable social,
educational, and economic futures for second-language learners becomes important. The
purpose of this study, therefore, was to investigate the efficacy of a kindergarten intervention
program aiming to improve second-language learners’ general language skills in their second
language.
Methods
Participants
Eligible participants were children for whom Norwegian was a second language and
who were born in 2006 and attended kindergartens in two suburban municipalities in
Norway. These municipalities were chosen because they were considered to be areas
representing the average socioeconomic status in Norway. However, children with an
immigrant background in Norway tend to have parents with a lower education and
earning rate than children of nonimmigrant parents, making them more disadvantaged
(see Bratsberg, Raaum, & Røed, 2011). This situation is also the case for the two
municipalities in the intervention.
To be included as a participant in the study, both parents had to have a first
language other than Norwegian. In all, 115 children (mean age D 66.18 months,
SD D 3.47) from 31 kindergartens participated in the study. Thus, all children in the
sample were in the final year of kindergarten, preceding entry to school. As is the case
in most European countries, second-language learners in Norway are highly heteroge-
neous in terms of their first languages; thus, there was no inclusion criterion based on
the children’s native languages. This situation resulted in a sample in which the partici-
pating children in the sample had different mother tongues from all over the world.
How long they had been in Norway when the study started also varied between the
participants. On this basis, the children varied in regard to how much exposure they
had to Norwegian. Randomizing at the individual level allowed for this mix in linguis-
tic backgrounds and varied experience with the second language between study partici-
pants. The children were assessed with Raven’s Coloured Progressive Matrices, sets A,
Ab, and B (Raven, Court, & Raven, 1990). The raw score mean for children in the
154 K. ROGDE ET AL.
intervention group was 14.43 (SD D 3.97), and the raw score mean for children in the
control group was 13.54 (SD D 3.65). Converted into percentile scores, using American
norms, both groups corresponded to a percentile of 50 for their age group. The flow of
participants through the trial is presented in Figure 1, in accordance with CONSORT
guidelines (Schulz, Altman, & Moher, 2010). Gender was equally distributed between
the intervention group (34 boys, 24 girls) and the control group (34 boys, 23 girls).
Ethical approval was obtained from the Norwegian Social Science Data Services, and
informed parental consent was obtained for each child in the study.
Norwegian kindergartners provide a childcare service that emphasizes informal learning.
The classrooms for this intervention consisted of children from ages 3 to 5, but only children
who were approximately 5 years of age (the year preceding school entrance) were selected to
participate in the intervention. The kindergartners use a play-based approach to learning
and emphasize the development of social skills. The program in the last year of kindergarten
also consists of school activities that are preparatory for reading, writing, and arithmetic.
Compulsory primary school begins the year that a child turns 6. In this study, kindergarten
teachers were responsible for the implementation of the intervention program. Employment
as a kindergarten teacher requires, at a minimum, a bachelor’s degree in kindergarten
Taught Vocabulary
Taught vocabulary was measured using a definition task to assess knowledge of words that
had explicitly been taught in the intervention. The test included a random sample of 28
words from a list of 56 taught words. The children were presented with a single word and
asked to provide an explanation of what the word meant. The response to each word was
scored on a scale of 0 through 3 points. For example, for the word “balance,” one point was
awarded for a demonstration or simple example (e.g., a demonstration of standing on one
leg), two points were awarded for a good example or explanation (e.g., “We must keep our
balance so that we don’t fall”), and three points were awarded for a synonym or a definition
(e.g., “equilibrium”). The internal consistency for the taught vocabulary test, as measured by
Cronbach’s alpha, was .76, .88, and .87 at time points t1, t2, and t3, respectively.
Narrative Skills
The Renfrew Bus Story test (Renfrew, 1997) was used to assess narrative skills. In this
test, the child is told a story, followed by instructions to retell it. The children’s retell-
ings were transcribed verbatim and assigned scores based on the vocabulary/key words
and story structure. The scoring system of the Norwegian version has been translated
from English to Norwegian for research purposes by researchers at the Department of
Special Needs Education at the University of Oslo. Cronbach’s alphas were .80, .82,
and .79 at time points t1, t2, and t3, respectively.
Intervention
The children were selected from 31 kindergartens and 55 different kindergarten classes. On
average, each class consisted of 20 children and approximately four adults (two teachers and
two assistants). The children were randomized on an individual level to a treatment or a
control condition. The participating children were located in 55 different classrooms.
Among these, 14 classes ended up having both children receiving the intervention and
children distributed to the control group. In these cases, preferably, the teachers from other
classes were responsible for the intervention. If this was not possible, to avoid problems
related to treatment diffusion, the teachers who participated in the intervention had thor-
ough instructions about the dangers of diffusion. These instructions included information
about the importance of not spreading the materials (e.g., books, worksheets) used in the
intervention. The teachers were also asked not to apply the dialogic reading technique when
reading for the children in the control group, and, preferably, get a teacher not involved in
the study to read for these children.
The intervention condition involved 54 instruction sessions that were taught three days
per week over a total of 18 weeks; each week, the instruction program involved two 40- to
45-minute small-group sessions and one 10-minute individual session for each child, given
in the following order: group session—individual session—group session (see Appendix A
in the online supplemental material). The intervention children also participated in the ordi-
nary kindergarten program but left the group for the 100 minutes per week that the
intervention lasted. Using a cost-effective approach, the program was implemented by kin-
dergarten teachers who received two days of training prior to the intervention. The control
GENERAL LANGUAGE INSTRUCTION 157
group followed a business-as-usual regime in which the children received the ordinary
kindergarten program, meaning that they followed the applicable county’s regular practice
for second-language learners. The regular practice was also based on reading and language
activities but was less structured and in larger groups than our intervention program. Addi-
tionally, there was less time with direct adult supervision in the regular practice group. How-
ever, note that each class had four adults, so even if a small-group child received an
intervention outside the group with one of the teachers, there were other teachers and adults
who could give sufficient attention to the control group.
The instruction program aimed to foster important skills within the general language
domain and language comprehension (i.e., phonological skills, letter knowledge, and print
awareness was not included in the sessions). Each of the kindergartners received all the nec-
essary material in advance of the intervention start-up so that they could familiarize them-
selves with the material. All teaching was done in line with a detailed teacher lesson script
(see Appendix B1 and B2 in the online supplemental material). The language teaching tasks
included classification of words and concepts, listening activities, exercises related to gram-
matical knowledge, story structuring, and story sequencing. The material was taken from
different sources. Large parts of the material used to teach narrative skills and grammar
were taken from the Black Sheep Press (http://www.blacksheeppress.co.uk/), recommended
by Carroll, Bowyer-Crane, Duff, Hulme, and Snowling (2011), which produces illustrated
worksheets and assessments to assist the development of children’s speech and language.
Several of the activity tasks used in the direct vocabulary instruction and concept classifying
were taken from the book Category Cut-ups: Vocabulary Building, Classifying, and Describ-
ing Cut and Paste Activities by Jennifer Larson (2006). In addition, the material that was
used for grammar and vocabulary teaching was also delivered by Taskmaster (http://www.
taskmasteronline.co.uk/), a company that specializes in teaching materials fostering speech
and language.
In addition to structured language tasks, an important component of the program was
based on dialogic storybook reading (developed by Whitehurst et al., 1988). Dialogic reading
involves specific questioning techniques used by teachers that are directed to the child to get
him or her to develop more context-independent language and to expand his or her vocabu-
lary while jointly attending to a book. Dialogical reading was chosen because meta-analyses
have shown that it is more efficient than conventional reading aloud (Mol, Bus, De Jong, &
Smeets, 2008). The first 20–25 minutes of the first group session every week, during the
entire intervention period, was spent on dialogic reading. In these sessions, books selected
by the research group were read and discussed in line with the detailed teacher lesson script.
Embedded as a part of the interactive reading method in these sessions, three or four words
were selected from each book and directly taught to the children. In addition, these words
were also repeated in the individual session that followed later in the week (see Appendix
B2). The purpose of incorporating these words in the intervention program was to examine
whether the vocabulary training given in the context of storybook reading was effective.
The selection of these words was based on the concept of tier-2 words that are domain-
general, sophisticated labels for concepts with which young learners are already familiar
(Beck & McKeown, 2007; Biemiller, 2009). The rationale for selecting tier-2 words was that
these are more abstract words that children will not easily learn by themselves in a general
kindergarten setting. These words are, however, highly relevant for building a more abstract
language and for later school performance. These words were also selected on the basis of
158 K. ROGDE ET AL.
school textbooks. We did not expect the children to know these words prior to the interven-
tion. Notably, work with these specific words constituted only approximately 10% of the
program because the main aim of the program was to produce transfer effects to more global
language skills extending beyond the specific words trained in the intervention.
Treatment Fidelity
Kindergarten teachers met for a two-hour introduction meeting and two days of training
sessions before the intervention started. The training sessions included information about
the teaching procedures of the intervention program and practice in which the teachers
worked with the specific instructions and materials. To ensure that each component in the
program would be delivered in a comparable manner to all study participants, each session
in the instruction program included detailed instructions for the kindergarten teachers to
follow.
During the intervention, the research group strived to monitor the implementation of the
intervention program in several ways. The teachers were asked to create audio recordings
and maintain logs from each training session. The logs were used to note the children’s
attendance and respond to questions asking whether the session had been successfully com-
pleted as instructed in the instruction manual. After the first two weeks of the intervention,
the research group called all the kindergarteners for updates and collected audio recordings
from five of the kindergarteners to examine whether the sessions went as planned. Halfway
through the intervention, the research group held a gathered follow-up meeting with the
kindergarten teachers to discuss experiences and to answer questions about the program.
Entering the last month of the intervention program, the research group once more called
the kindergartners for updates and offered individual visits.
To assess the fidelity of program implementation after the intervention finished, a ran-
dom selection of 5% (59 sessions) of the total number of sessions across all the kindergart-
ners was checked to determine whether the components in the sessions had been delivered
or not and delivered as instructed. Among the selected sessions, 53 of the 59 sessions had
been conducted, yielding an implementation rate of 90%. Thus, there were six sessions that
had not been completed. These sessions were mainly omitted from the final one to three
weeks of the intervention. It may be that the last weeks of the intervention program were
placed too close to the end of the kindergarten year, making it too hectic to implement
the intervention program. Another explanation may be that the kindergarten teachers
became tired and less loyal to the intervention program at the end of the project. To investi-
gate whether the sessions had been delivered as prescribed in the program, audio recordings
from the selected sessions (i.e., the remaining 53 sessions that had been carried out) were lis-
tened to and compared against the guided instructions and the events reported in the logs.
For all these sessions, all the given activities had been conducted and the instruction for the
activities had been followed correctly. There was also compliance between the audio record-
ings and the information given in the logs.
Results
The mean values and standard deviations of the children’s raw scores for all measures at
each time point (t1, t2, and t3) are shown in Table 1. The intervention and control groups
showed no noteworthy differences in the pretest scores on any of the outcome measures.
GENERAL LANGUAGE INSTRUCTION 159
Table 1. Mean raw scores and standard deviations for the outcome measures at pretest (t1), posttest 1 (t2)
and posttest 2 (t3).
Intervention group Control group
M SD M SD
Age in months
Time 1 66.28 3.27 66.09 3.69
Time 2 71.48 3.38 70.73 3.80
Time 3 78.50 3.43 78.33 3.61
Taught vocabulary
Time 1 7.55 5.52 6.89 4.53
Time 2 17.04 10.45 12.18 6.69
Time 3 19.00 8.66 15.23 8.66
WPPSI/WISC vocabulary
Time 1 11.76 7.44 12.36 7.26
Time 2 16.57 7.28 13.59 7.58
Time 3 18.96 8.10 17.39 7.63
ITPA grammar
Time 1 8.21 4.64 7.48 4.66
Time 2 11.71 4.62 9.59 4.47
Time 3 13.55 4.42 12.36 4.71
BPVS II vocabulary
Time 1 46.79 16.77 46.66 15.44
Time 2 55.21 14.38 53.89 16.71
Time 3 66.80 14.44 62.75 15.85
TROG 2 grammar
Time 1 32.07 17.00 30.00 14.75
Time 2 41.66 17.18 40.42 16.19
Time 3 52.32 14.62 50.86 15.22
Bus Story narrative
Time 1 11.28 7.13 10.96 7.13
Time 2 18.14 8.22 13.75 7.82
Time 3 18.93 7.83 16.79 7.54
WPPSI/WISC: Wechsler Preschool and Primary Scale of Intelligence/Wechsler Intelligence Scale for Children; TROG: Test for
Reception of Grammar; British Picture Vocabulary Scale; BPVS: British Picture Vocabulary Scale; ITPA: Illinois Test of Psycho-
linguistic Abilities.
The analyses of effects were performed with Mplus, version 7.3 (Muthen & Muthen, 1998–
2014), using latent autoregressive models (structural equation models) with maximum-like-
lihood estimation to handle missing data. Huber-White correction was used to control for
kindergarten dependency in all analyses. These types of analyses were chosen over
MANCOVA because they estimate relations between the latent constructs without measure-
ment errors and are better able to both reflect and test our theoretically defined language fac-
tors. When a construct was represented with only one variable, this variable was divided into
two halves, each of which was used as an indicator for the latent construct, to avoid attenua-
tion due to measurement error.
Measurement Models
First, we tested whether the five measures of general language skills could be seen as reflecting
one overall language factor or whether they were better seen as reflecting two factors: expressive
(with the indicators expressive grammar, expressive vocabulary, narrative retelling) and recep-
tive (with the indicators receptive vocabulary and receptive grammar). This was done by testing
whether constraining the correlation between the expressive and receptive factor to 1 (i.e., a
one-factor solution) fitted the data significantly worse than merely freely estimating correlation
160 K. ROGDE ET AL.
between the two factors (i.e., a two-factor solution). Wald tests showed that one factor was suffi-
cient at Time 1, x2 (1) D 1.573, p D .231, and Time 3, x2 (1) D .930, p D .335, but not at Time 2,
when the two-factor solution fitted the data significantly better than the one-factor solution,
x2 (1) D 7.731, p D .005, albeit with a high correlation between the two factors (.85). Because of
this difference at the first posttest, we chose to analyze the effects of the language training sepa-
rately for the two factors.
Next, we tested for longitudinal measurement invariance to determine whether the differ-
ent factors could be defined in the same way at all three time points. We found full scalar
invariance (factor loading and intercepts) for taught vocabulary, Dx2(4) D 4.919, p D .296,
and receptive language, Dx2(4) D 4.919, p D .296. For expressive language, we found full
metric invariance (factor loadings), Dx2(4) D 3.781, p D .436, and partial scalar invariance,
Dx2(3) D 5.712, p D .149 (difference between the models with metric and partial scalar
invariance). In the case of the expressive factor, the intercept of expressive vocabulary was
allowed to vary at time point 1. Thus, the latent variables could be defined in a similar way
at the three time points. These final models fitted the data well: taught vocabulary, x2 (10) D
6.431, p D .778, RMSEA D .000 (90% CI D .00–.069), CFI D 1.00, SRMS D .045, receptive
language, x2 (7) D 8.205, p D .315, RMSEA D .039 (90% CI D .00–.125), CFI D .997,
SRMS D .055, and expressive language, x2 (26) D 35.278, p D .106, RMSEA D .056 (90%
CI D .00–.098), CFI D .981, SRMS D .042. It should also be noted that a nonsignificant Hey-
wood case was found in the model of receptive language: the correlation between the latent
factor at times 1 and 2 was 1.03. However, because this correlation did not differ from 1
(p D .728), we have statistical support for it being sampling fluctuations and it can be
ignored.
Altogether, these results indicate that the various concepts can be interpreted in a similar
fashion over time as long as we divide language comprehension into expressive and receptive
factors.
language model, x2 (33) D 37.80, p D .260, RMSEA D .036 (90% CI D .000–.080), CFI D .991,
SRMS D .041, and for the receptive language model, x2 (12) D 13.75, p D .317, RMSEA D .036
(90% CI D .000–.105), CFI D .996, SRMS D .054.
Discussion
This RCT evaluated an 18-week intervention program implemented in kindergartens
with the aim of improving second-language learners’ general language skills in the sec-
ond language. The findings showed clear immediate effects (posttest 1) of the instruc-
tion on the custom measure of taught vocabulary. This finding is consistent with the
GENERAL LANGUAGE INSTRUCTION 163
findings of Roberts and Neal (2004) and Collins (2010) and supports the idea that rich
explanations of words through storybook reading are an efficient and age-appropriate
tool with which to teach younger second-language learners the meanings of target
words. These results are also in line with those of many other studies of language inter-
ventions on first language learners—namely, that near-transfer effects can be obtained
on criterion measures similar to the tasks that are trained (Elleman et al., 2009, Marulis
& Neuman, 2010). Although these results are not very surprising, they are promising in
that they show that vocabulary instruction of words to prepare second-language learn-
ers for the language encountered in textbooks is effective.
Given what has been learned in earlier research (Marulis & Neuman, 2010; National Insti-
tute of Child Health and Human Development, 2000; Stahl & Fairbanks 1986), we can
assume that including explicit instruction of word meanings, multiple exposures, and allow-
ing the children to interact with the words that are discussed are factors that may have con-
tributed to the effect. Notably, the tasks in our intervention are not based on children’s
rehearsal of the definitions of the specific words selected for the program. Rather, the words
are embedded in the program using book sessions and are presented in a natural setting.
The focus on presenting the word in context rather than rehearsal definitions is likely to
explain why the present study found somewhat smaller effects on the trained words than
what has been found in other studies (see Elleman et al., 2009).
Regarding far-transfer effects, the intervention produced immediate effects on expressive
language skills. Thus, our language instruction program was able to improve expressive lan-
guage skills beyond the specific components taught in the instruction (e.g., target words). In
the light of the relative stability found among children in the development of language skills
(Bornstein et al., 2014; Klem et al., 2014; Melby-Lervag et al., 2012; Storch & Whitehurst,
2002), these results are encouraging. The results highlight the possibility of enhancing sec-
ond-language learners’ expressive language skills in the second language prior to school
entry. It is thus clear that although general language ability is a strongly preserved individual
characteristic in which the rank order between the children is highly stable (Bornstein et al.,
2014), these skills are malleable through interventions. However, note that Bornstein et al.
(2014) found smaller stability coefficients in younger children than in older children, mean-
ing that early intervention to improve language skills might be recommended.
In our analyses, we found support for a unitary language construct at baseline, but after
the training, a two-factor model with expressive and receptive language as two correlated
constructs fit the data best. Additionally, the effects from training were related to standard-
ized measures of expressive language skills rather than the language factor in general. A rele-
vant question, then, is whether the fact that we obtained differential effects from the training
can be seen as refuting the theory of a global language factor. In our view, the fact that a
two-factor solution fit the data best at Time 2 reflects that although a one-factor solution fit
the construct best at Time 1, the construct has different facets (e.g., expressive vs. receptive)
and that these facets can be differentially affected by training. Because the training led to
effects on expressive language measures and not the receptive measures, a two-factor solu-
tion turned out to be the preferred fitting model after the training. At follow-up, however,
the effects from training faded a bit and the two-factor solution was no longer different from
the one-factor solution. Thus, our findings are not inconsistent with the theory of a general
global language factor, but our findings provide support that different interventions can
address different facets of this general construct.
164 K. ROGDE ET AL.
Considering the differential effects on receptive and expressive measures, these effects
may be explained by the fact that the majority of the tasks and games in our intervention
involved oral responses and hence fostered expressive language more than receptive lan-
guage skills. Furthermore, in light of the results showing that the challenges of L2 learners
are particularly severe on expressive skills in the second language (Gibson et al., 2012), the
children in our study may have had a greater potential for change in expressive skills. If we
compare our results to other findings, the lack of effects on receptive language measures has
also been the case in other studies on dialogic reading using a similar receptive vocabulary
measure (PPVT-R; Dunn & Dunn, 1981; Lonigan & Whitehurst, 1998; Whitehurst et al.,
1994; Whitehurst et al., 1988). However, book reading studies in which receptive vocabulary
has been assessed by more sensitive measures (using vocabulary introduced in the book)
have demonstrated more positive results (Senechal, 1997; Senechal & Monker, 1995). It is
possible that standardized measures of receptive vocabulary are less sensitive to vocabulary
growth, making it difficult to identify changes. However, as shown by Senechal (1997), dif-
ferential effects on expressive vs. receptive language may also be due to didactic techniques
used by the adult in storybook reading.
Our finding of an effect on expressive language is similar to that of Farver et al. (2009), in
which English-language learners in preschool increased their expressive vocabulary skills after
small-group instruction spanning 21 weeks. In contrast, Farver et al.’s (2009) study produced
effects on a receptive vocabulary measure similar to the one used in our study. Unfortunately,
Farver et al.’s (2009) study did not include follow-up tests to check for possible retention
effects. Clearly, further longitudinal developmental studies and experimental studies are
needed to develop a better understanding of the components of language instruction and fea-
tures in the implementation of interventions that produce gains in receptive language skills.
The results from the seven-month follow-up test (posttest 2) indicate that the effect on
the taught vocabulary was maintained. There was also a follow-up effect on the expressive
language construct, albeit reduced in size from the first posttest. We can only speculate about
possible explanations for why the effect on the expressive language construct decreased at
follow-up. One explanation may be that the follow-up effects would more likely have been
maintained had the instruction been continued. Another possible explanation may be that
all the participating children were challenged to make use of language skills and were more
exposed to activities fostering expressive language skills in the school context after the inter-
vention ended (i.e., the time between the first and second posttest). This situation may have
led the children in the control group to close the gap relative to the children in the treatment
group. However, the observed “fade-out” of gains in this intervention is in line with that
found in other studies of intervention effects in early childhood (see Barnett, 2011;
Guralnick, 2005 for a discussion of this). Such results speak to the need to strengthen our
knowledge of the effectiveness of interventions, and future studies should value a more long-
term perspective. Perhaps it could be useful to offer shorter blocks of interventions in
schools accompanied by repeated measures of effects, rather than the more traditional large
block of intervention with immediate effect and follow-up measures.
The treatment effects in our study did not differ as a function of initial language status.
This finding implies that second-language learners can benefit from general language
instruction even when they begin from a low initial point. The intervention program ensured
that each child received opportunities to engage in the activities. In addition, the kindergar-
ten teachers were instructed to adapt their discussions to the children and to facilitate the
GENERAL LANGUAGE INSTRUCTION 165
activities using concrete activities and repetitions when necessary. This approach may have
allowed the children to benefit equally from the instruction, regardless of their language
competence at the beginning of the study.
Finally, our findings demonstrate that it is possible to support second-language learners’
general language skills in the second language using a cost-effective approach to instruction
in kindergartens. However, interpreting effect sizes to provide meaningful descriptions
for practical importance is not straightforward (see Cooper, 2008 for a discussion). A princi-
ple practiced by the What Works Clearinghouse (U.S. Department of Education, 2014) and
Promising Practices Network (2014) is to consider an effect size of 0.25 standard deviations
or larger as “substantively important.” According to this standard, the results of this study
indicate that general language instruction for young second-language learners is a matter of
both immediate and longer term practical interest.
Acknowledgments
We would like to thank the school district municipalities for their assistance throughout the project.
We would also like to thank the children, parents, preschool teachers, and research assistants who par-
ticipated in the study.
Funding
This study was funded by the Research Council of Norway, Utdanning 2020, Grant No.: 203335.
166 K. ROGDE ET AL.
ARTICLE HISTORY
Received 10 December 2014
Revised 11 March 2016
Accepted 21 March 2016
EDITORS
This article was reviewed and accepted under the editorship of Carol McDonald Connor and Spyros
Konstantopoulos.
References
Barnett, S. M., & Ceci, S. J. (2002). When and where do we apply what we learn? A taxonomy for far
transfer. Psychological Bulletin, 128, 612–637. doi:10.1037/0033-2909.128.4.612
Barnett, W. S. (2011). Effectiveness of early educational intervention. Science, 333, 975–978.
doi:10.1126/science.1204534
Beck, I. L., & McKeown, M. G. (2007). Increasing young low-income children’s oral vocabulary reper-
toires through rich and focused instruction. Elementary School Journal, 107, 251–271. doi:10.1086/
511706
Beitchman, J. H., Wilson, B., Johnson, C. J., Atkinson, L., Young, A., Adlaf, E.,… Douglas, L. (2001).
Fourteen-year follow-up of speech/language-impaired and control children: Psychiatric outcome.
Journal of the American Academy of Child & Adolescent Psychiatry, 40, 75–82. doi:10.1097/
00004583-200101000-00019
Biemiller, A. (2009). Words worth teaching. Columbus, OH: SRA/McGraw-Hill
Bishop, D. (2003). Test for reception of grammar. Version 2. Trog-2 manual (Norwegian translating
and norms: Lyster, S.-A. H. & Horn, E., 2009). Stockholm, Sweden: Pearson
Bornstein, M. H., Hahn, C. S., Putnick, D. L., & Suwalsky, J. T. (2014). Stability of core language skill
from early childhood to adolescence: A latent variable approach. Child Development, 5, 1346–1356.
doi:10.1111/cdev.12192
Bowyer-Crane, C., Snowling, M. J., Duff, F. J., Fieldsend, E., Carroll, J. M., Miles, J., … Hulme, C.
(2008). Improving early language and literacy skills: Differential effects of an oral language versus a
phonology with reading intervention. Journal of Child Psychology and Psychiatry, 49, 422–432.
doi:10.1111/j.1469-7610.2007.01849.x
Bransford, J. D., & Schwartz, D. L. (1999). Rethinking transfer: A simple proposal with multiple impli-
cations. In A. Iran-Nejad & P. D. Pearson (Eds.), Review of research in education (Vol. 24, pp. 61–
100). Washington, DC: American Educational Research Association
Bratsberg, B., Raaum, O., & Røed, K. (2011). Educating children of immigrants: Closing the gap in Nor-
wegian schools (Discussion Paper Series, IZA DP no. 6138). Retrieved from http://www.iza.org/en/
webcontent/publications/papers/viewAbstract?dp_idD6138
Brown, T. A. (2006). Confirmatory factor analysis for applied research. New York, NY: The Guilford
Press
Carraher, D., & Schliemann, A. (2002). The transfer dilemma. Journal of the Learning Sciences, 11,
1–24. doi:10.1207/S15327809JLS1101_1
Carroll, J. M., Bowyer-Crane, C., Duff, F. J., Hulme, C., & Snowling, M. J. (2011). Developing language
and literacy: Effective intervention in the early years. Oxford, UK: Wiley-Blackwell.
Child Trends Databank. (2011). Child Trends’ calculations of U.S. Census Bureau, school enrollment—
Social and economic characteristics of students: October 2005 (Detailed Tables: Table 1). Retrieved
from http://www.census.gov/population/www/socdemo/school/cps2005.html
Cohen, N. J. (2001). Language impairment and psychopathology in infants, children, and adolescents.
New York, NY: Sage Publications
GENERAL LANGUAGE INSTRUCTION 167
Colledge, E., Bishop, D. V., Koeppen-Schomerus, G., Price, T. S., Happe, F. G., Eley, T. C.,… Plomin,
R. (2002). The structure of language abilities at 4 years: A twin study. Developmental Psychology,
38, 749–757. doi:10.1037/0012-1649.38.5.749
Collins, M. F. (2010). ELL preschoolers’ English vocabulary acquisition from storybook reading. Early
Childhood Research Quarterly, 25, 84–97. doi:10.1016/j.ecresq.2009.07.009
Cooper, H. (2008). The search for meaningful ways to express the effects of interventions. Child Devel-
opment Perspectives, 2, 181–186. doi:10.1111/j.1750-8606.2008.00063.x
Droop, M., & Verhoeven, L. (2003). Language proficiency and reading ability in first- and second-lan-
guage learners. Reading Research Quarterly, 38, 78–103. doi:10.1598/RRQ.38.1.4
Dunn, L. M., & Dunn, L. M. (1981). Peabody Picture Vocabulary Test—Revised. Circle Pines, MN:
American Guidance Service
Dunn, L. M., Dunn, L. M., Whetton, C., & Burley, J. (1997). British picture vocabulary scale II.
Windsor, England: National Foundation for Educational Research-Nelson
Elleman, A. M., Lindo, E. J., Morphy, P., & Compton, D. L. (2009). The impact of vocabulary instruc-
tion on passage-level comprehension of school-age children: A meta-analysis. Journal of Research
on Educational Effectiveness, 2, 1–44. doi:10.1080/19345740802539200
Farver, J. A., Lonigan, C. J., & Eppe, S. (2009). Effective early literacy skill development for young
Spanish-speaking English language learners: An experimental study of two methods. Child Devel-
opment, 80, 703–719. doi:10.1111/j.1467-8624.2009.01292.x
Foorman, B., Herrera, S., Petscher, Y., Mitchell, A., & Truckenmiller, A. (2015). The structure of oral
language and reading and their relation to comprehension in kindergarten through grade 2. Read-
ing & Writing, 28, 655–681
Foorman, B., Koon, S., Petscher, Y., Mitchell, A., & Truckenmiller, A. (2015). Examining general and
specific factors in the dimensionality of oral language and reading in 4th–10th grades. Journal of
Educational Psychology, 107, 884–899. doi:10.1037/edu0000026
Fricke, S., Bowyer-Crane, C., Haley, A. J., Hulme, C., & Snowling, M. J. (2013). Efficacy of language
intervention in the early years. Journal of Child Psychology and Psychiatry, 54, 280–290.
doi:10.1111/jcpp.12010
Gibson, T. A., Oller, D. K., Jarmulowicz, L., & Ethington, C. A. (2012). The receptive–expressive gap in
the vocabulary of young second-language learners: Robustness and possible mechanisms. Bilingual-
ism: Language and Cognition, 15(1), 102–116. doi:10.1017/S1366728910000490
Guralnick, M. J. (2005). An overview of the developmental systems model for early intervention. In M.
J. Guralnick (Ed.), The developmental systems approach to early intervention (pp. 3–28). Baltimore,
MD: Paul H. Brookes.
Johnson, C. J., Beitchman, J. H., Young, A., Escobar, M., Atkinson, L., Wilson, B.,… Wang, M. (1999).
Fourteen-year follow-up of children with and without speech/language impairments: Speech/lan-
guage stability and outcomes. Journal of Speech, Language and Hearing Research, 42, 744–760.
doi:10.1044/jslhr.4203.744
Kieffer, M. J., & Vukovic, R. K. (2013). Growth in reading-related skills of language minority learners
and their classmates: More evidence for early identification and intervention. Reading and Writing,
26, 1159–1194. doi:10.1007/s11145-012-9410-7
Kirk, S. A., McCarthy, J. J., & Kirk, W. D. (1968). Illinois test of psycholinguistic abilities. Urbana: Uni-
versity of Illinois Press
Klein, A. G., & Moosbrugger, H. (2000). Maximum likelihood estimation of latent interaction effects
with the LMS method. Psychometrika, 65, 457–474
Klem, M., Melby-Lervag, M., Hagtvet, B., Lyster, S.-A. H., Hulme, C., & Gustafsson, J. E. (2014). Sen-
tence repetition is a measure of children’s language skills rather than working memory limitations.
Developmental Science, 18, 146–154. doi:10.1111/desc.12202
Landry, S. H., Swank, P. R., Anthony, J. L., & Assel, M. A. (2011). An experimental study evaluating
professional development activities within a state funded pre-kindergarten program. Reading and
Writing, 24, 971–1010
Larson, J. (2006). Category cut-ups: Vocabulary building, classifying, and describing cut and paste activ-
ities. Greenville, SC: Super Duper Publications