Content uploaded by Florina Erbeli
Author content
All content in this area was uploaded by Florina Erbeli on Jun 27, 2021
Content may be subject to copyright.
EFFECTS OF INDEPENDENT READING 1
To cite this article:
Erbeli, F., & Rice, M. (2021). Examining the effects of silent independent reading on reading
outcomes: A narrative synthesis review from 2000 to 2020. Reading & Writing Quarterly.
https://doi.org/10.1080/10573569.2021.1944830
Running head: EFFECTS OF INDEPENDENT READING
Examining the effects of silent independent reading on reading outcomes: A narrative synthesis
review from 2000 to 2020
Florina Erbeli, Marianne Rice
Author Notes
Florina Erbeli, Department of Educational Psychology, Texas A&M University; ORCID: 0000-
0002-4955-3170
Marianne Rice, Department of Educational Psychology, Texas A&M University; ORCID: 0000-
0001-8935-4734
Conflict of interest statement: no conflicts declared.
Correspondence concerning this article should be addressed to Dr. Florina Erbeli,
Department of Educational Psychology, 4225 Texas A&M University, College Station, TX,
77843, ph. 979-845-3521, e-mail: erbeli@tamu.edu.
EFFECTS OF INDEPENDENT READING 2
Abstract
Encouraging children to read extensively has been a widely recommended approach to
developing reading. The National Reading Panel (NRP, 2000) published a review study reporting
inconclusive findings regarding the benefits of such an approach. In this systematic narrative
synthesis review, we provided an update and an extension of the NRP’s review. We examined
the effects of silent independent reading practices on reading outcomes for students in Grades K
through 12, reviewing experimental and quasi-experimental studies between 2000 and 2020. We
also incorporated a quality evaluation of primary studies. A systematic search of peer-reviewed
articles was conducted, using identical procedures as in the NRP review. Our results from 14
primary studies comprising 5,522 participants in the treatment group and 4,966 in the control
group alluded to no meaningful beneficial effects of independent reading on reading outcomes.
However, due to a lack of primary studies adhering to the highest quality standards and
implementation, it is impossible to determine whether such a result is universal or whether there
might be conditions under which independent reading could be effective.
EFFECTS OF INDEPENDENT READING 3
Examining the effects of silent independent reading on reading outcomes: A
narrative synthesis review from 2000 to 2020
It is intuitively appealing to believe that silent independent reading will have a positive
impact on a range of reading outcomes. In fact, there is a widespread acceptance of the idea
among theorists, researchers, and practitioners alike that encouraging children and adolescents to
read extensively on their own translates into better reading achievement outcomes. However,
does extended reading in and of itself make children and adolescents better readers? The answer
to this question is not straightforward. In the reading fluency subsection of the National Reading
Panel’s report (NRP, 2000), researchers concluded that experimental research has not yet clearly
demonstrated such a causal direction of an effect. In other words, the NRP neither accepted nor
rejected the position that extensive silent independent reading is an efficient way to increase
reading outcomes (NRP, 2000, p. 3-27). The main argument for such a conclusion was based on
the appraisal of the primary experimental studies. Primary studies showed serious weaknesses
and limitations in instructional procedures and methodological designs that they implemented
(NRP, 2000, p. 3-27). Since the NRP’s report release in 2000, a set of quality standards has been
recommended, which experimental studies must implement (What Works Clearinghouse; WWC,
2003). In addition, systematic review report standards have been updated and revised (e.g.,
Appelbaum et al., 2018). The latest directions to advance rigorous research in the fields of
psychology and education (Appelbaum et al., 2018; WWC, 2003) along with NRP’s agnostic
conclusion about the impact of independent reading on different reading outcomes have
prompted us to undertake this work. The present study is a systematic narrative synthesis review
aimed at reporting the effectiveness of silent independent reading practices on reading related
outcomes and reading achievement in children and adolescents in grades K through 12. We
EFFECTS OF INDEPENDENT READING 4
apply rigorous methodological reporting standards for synthesis reviews as well as evaluate the
quality of primary studies to summarize the current state of the art in this area from 2000 to
2020. Given our goal, our study can be viewed as an update of the NRP’s (2000) reading fluency
synthesis review on independent reading, addressing its methodological and reporting
limitations.
Extensive silent independent reading is defined here identically as in the NRP (2000)
report. Students read silently and individually on their own with no monitoring and with little or
no specific feedback (NRP, 2000, p. 3-21). In most cases, students select their own reading
materials, and there is no discussion or written assignment tied to the reading. The NRP’s (2000)
and our definition rules out any examples of practices, in which independent reading is
accompanied by teacher interaction and/or incorporates any additional reading instruction aside
from silent independent reading. In addition, the definition excludes examples of practices that
include teacher graded quizzes and/or high-stakes testing accountability through completed book
response assignments. The NRP report (2000, p. 3-22) lists many synonyms for extensive silent
independent reading, which will also be used here. For instance, free time silent reading,
recreational reading, voluntary reading are all examples of an identical concept. (See Boolean
search terms in Supplemental Materials for more synonyms.)
Even though the NRP’s (2000) reading fluency synthesis review on independent reading
reported inconclusive findings about silent independent reading interventions being a scalable
instructional approach for improving reading outcomes, the theoretical ground behind an idea of
reading practice impacting reading is not implausible. LaBerge and Samuels’ (1974) theory of
automatic information processing in reading and Perfetti’s (1985) verbal efficiency hypothesis
state that students who are expending their cognitive resources on decoding words have little or
EFFECTS OF INDEPENDENT READING 5
no cognitive resources left for comprehending what they read. With ample practice and repeated
exposure, the reader attains automaticity in word recognition, freeing resources to perform two
reading tasks at once, decoding and comprehending text. In this sense, continued reading practice
is beneficial because word recognition becomes automatic, which, in turn, promotes reading
comprehension. Moreover, Perfetti’s theory (1985) posits that not only word recognition and
decoding, but also higher level reading processes integral for reading comprehension, such as
integrating propositions, using cognitive and meta-cognitive strategies, and activating
background schemas, can be automatized through extended reading practice. In sum, this means
that the importance of automatic and, as such, effective reading processes resulting from reading
practice cannot be disputed to be a valuable prerequisite to skilled reading. However, this
conclusion begs the question about the conditions under which reading practices might facilitate
reading outcomes (e.g., how much silent reading practice is most efficient for which students).
As previously mentioned, the NRP (2000) report could not unambiguously answer this
question. The panel systematically reviewed 14 experimental and quasi-experimental studies of
variable quality to say that any sort of positive or negative conclusion about the value of
independent silent reading practice was not feasible due to serious methodological or reporting
flaws in the studies. The panel debated about at least three issues that undermined the
inconclusive results of those 14 studies (NRP, 2000, p. 3-24). First, researchers did not measure
nor demonstrate adequate baseline equivalence among participants on the key characteristic –
amount of reading – before the intervention. Comparing treatment independent reading routines
against typical reading practices was meaningless since any equivalence at baseline between the
treatment and control groups on a studied trait could not be deduced from the studies. Second, a
few of the studies reviewed monitored the amount of silent independent reading students did in
EFFECTS OF INDEPENDENT READING 6
the treatment conditions and only one kept track of the control student reading. Hence, it could
not be concluded with certainty whether the interventions led to more reading or just displaced
independent reading students might have done otherwise (NPR, 2000, p. 3-27). The panel called
to an importance of comparing the amount of silent independent reading against routines in
which students actually read less. Third, studies that compared reading outcomes of students in
independent reading interventions versus controls during a regular school day did not isolate the
effects of the intervention on student outcomes from regular instructional effects. In some of the
identified studies, students in the intervention groups received considerably more instructional
time in reading than their control peers, confounding a potential intervention effect with the
effects of the additional instructional time in reading. Studies would need to consider the
duration of an intervention separate from the amount of instructional time in reading by both the
experimental and control groups. Taken together, these limitations precluded the NRP (2000)
panel from demonstrating presence or absence of support for the use of independent reading
practices to facilitate reading outcomes. The majority of included studies did not adopt a set of
rigorous research methodological standards that would permit a unified robust conclusion and
sustain a claim of effectiveness of independent reading practices on reading outcomes.
Because of the recommendation by the NRP, many schools completely limited
independent reading practices in classrooms immediately after the NRP’s report release in 2000
(Hiebert & Reutzel, 2014). However, in recent years independent reading in schools has been on
the rise again, with many schools implementing popular independent reading programs, such as
Accelerated Reader (Hiebert & Reutzel, 2014). Hence, more studies have been conducted to
determine whether the findings from the NRP’s reading fluency synthesis review on independent
reading continue to be valid. Research from individual studies has yielded mixed findings
EFFECTS OF INDEPENDENT READING 7
(Merga, 2013). What is more concerning, however, is that studies have continued to lack
adherence to highest quality standards much the same way as the NRP established in 2000. For
example, the U.S. Department of Education, WWC division, conducted a review of 318 studies
on the effectiveness of Accelerated Reader program on adolescent reading outcomes. They found
that only two studies (one of which was a dissertation) were randomized control trials that met
standards with reservations (WWC, 2010). All other studies were of questionable quality and
lacked rigorous methodological implementation and validity (WWC, 2010).
Continuation of inconclusive findings first by the NRP’s report and later by the WWC
reports has led us to update, complement, and extend their work, while examining an identical
question of a potential impact of independent reading on reading outcomes. Specifically, we
provide a systematic narrative synthesis review for studies conducted since NRP’s synthesis
review release in 2000, using the same criteria as the panel to synthesize an updated corpus of
literature. Moreover, we extend the panel’s report by incorporating a quality evaluation of the
included studies. Our intention for adding the application of quality indicators is to evaluate the
merits of reviewed studies and add credibility to the findings. We, thereby, adhere to the
recommendations by the WWC (2003) on a set of quality standards intervention studies must
implement and to the systematic review reporting standards, developed by the American
Psychological Association (Appelbaum et al., 2018). No protocol was published for this review.
Method
Information Sources and Study Selection
A three-step process was used to identify studies that examined the effects of independent
reading on reading outcomes. In the first step, we conducted a comprehensive computerized
search of PsycINFO, ERIC, Academic Search Ultimate, and PsycArticles databases using
EFFECTS OF INDEPENDENT READING 8
Boolean search terms in accordance with the terminology used in the NRP report (2000, p. 3-22).
The Boolean search terms can be found in Supplemental Materials. Since the purpose was to
review studies completed after the NRP report (2000) was released, the search was limited to
publications between the years 2000-2020. The search was conducted on May 21, 2020. No
changes to the search and prespecified inclusion and exclusion criteria have been made since
then. The initial search yielded a total of 28,630 studies (PsycINFO 10,125; ERIC 7,712;
Academic Search Ultimate 10,472; PsycArticles 321). After removing 12,208 duplicates, the
pool consisted of 16,422 studies for primary screening. The first author (an assistant professor in
educational psychology department) and the second author (a doctoral student in educational
psychology department) reviewed the studies by reading titles and abstracts on the platform
Rayyan, sorting them into three categories (“yes”, “no”, “maybe”) based on the inclusion and
exclusion criteria described below. The studies in the “yes” and “maybe” categories were more
closely scrutinized to ensure they met the inclusion criteria and were reallocated to a different
category as needed. In the primary screening, 16,398 studies were excluded for various reasons
(see Figure 1). This screening resulted in 24 studies, which were eligible for secondary
screening.
After completion of the computerized search, we conducted the second step of the
literature search. We reviewed the WWC Intervention Reports on various independent reading
interventions (e.g., Accelerated Reader) to ensure that all relevant studies had been identified in
the computerized search. From this search method, two additional journal articles were identified
(i.e., Nunnery & Ross, 2007; Shannon et. al, 2015) and included in the study pool for secondary
screening.
EFFECTS OF INDEPENDENT READING 9
Lastly, the third step, an ancestry search (Cooper, 2017), was conducted examining the
reference lists of included articles to search for additional potential studies, which were not
identified in the first two literature search steps. Three additional studies for secondary screening
were identified through this method (i.e., Kim & White, 2008; Pavonetti et al., 2003; Reutzel et
al., 2012).
For secondary screening, full articles of the resulting 29 studies were obtained and
reviewed carefully for eligibility. Of the 29 studies, 15 were excluded from the final synthesis for
various reasons. The flowchart presented in Figure 1 illustrates the literature search procedure
described above.
Inclusion and Exclusion Criteria
The inclusion and exclusion criteria for the eligibility of each study were similar to those
used in the NRP report (2000, p. 3-23). 1. The study had to be an experimental or quasi-
experimental report that considered the effect of encouraging students to read independently on
reading related outcomes and reading achievement. Studies not reporting numerical data on
outcome measures to compute effect sizes were still included since we did not conduct a meta-
analysis, which would have required complete numerical data to calculate a pooled effect size.
Studies failing to include a control group and studies analyzing the association between
voluntary reading and reading outcomes were excluded. In addition, we excluded studies not
containing sufficient information to judge eligibility in terms of study design. Lastly, studies not
including at least one reading achievement outcome as a dependent variable were excluded. 2.
The study had to focus on English reading education, conducted with children and adolescents in
grades K through 12. Studies conducted on college students or other adult populations were
excluded. 3. The study itself had to have appeared in a refereed journal. Theses, dissertations,
EFFECTS OF INDEPENDENT READING 10
evaluation reports, conference presentations and submissions, and any unpublished manuscripts
were excluded. 4. The study had to have been carried out with English language reading.
Data Collection and Methods for Assessing Risk to Internal Validity
To extract data about each study, we adapted a coding sheet developed within the
Meadows Center for Preventing Educational Risk at the University of Texas at Austin (Vaughn
et al., 2014). The coding sheet includes critical categories identified in the systematic review
process of the WWC Study Review Guide (2020). The following categories were included in the
codebook: general study characteristics, participants information (e.g., grade of participants),
design information (e.g., assignment of participants for intervention), treatment and comparison
groups (e.g., duration of intervention, description of the treatment and comparison conditions),
clarity of causal inference (e.g., differential attrition, equivalence at baseline), general findings,
precision of outcome (e.g., statistical assumption of independence), measures (e.g., types of
measures and their reliability and validity estimates), and effect sizes (e.g., direction of effects).
We created a second code sheet that included Gersten et al.’s (2005) and the WWC’s
(2020) quality indicators for experimental or quasi-experimental group-design studies that were
not included in the first code sheet but were relevant to our research question. Quality indicators
were used to assess risk to internal validity and show the degree to which reviewed studies were
conducted using high-quality designs. Quality indicators in the second code sheet, which
complemented categories from the first code sheet, were the following: describing participants
(e.g., baseline equivalence on the key characteristic – amount of independent reading – before
the intervention), implementation of the intervention and description of comparison conditions
(e.g., description and assessment of fidelity of implementation, description of the nature of
EFFECTS OF INDEPENDENT READING 11
services provided in comparison conditions), and outcome measures (e.g., use of measures
closely aligned with the intervention and measures of generalized performance).
Before starting coding, the second author was trained by the first author on the use and
interpretation of categories from the codebooks. The reliability of coding was established at the
start of coding by double coding two sample studies and verifying the coding. Next, the first and
second authors coded all included studies independently. The mean raw agreement across the
total category matrix was 97.6%. Disagreements in coding between raters were resolved by
consulting the original article. Once coding had been completed, study information and quality
appraisals were summarized in a tabular and graphical form. A figure presenting quality
evaluations was created using the risk-of-bias visualization tool in the program R, version 3.6.1
(McGuinness & Higgins, 2021).
Methods of Synthesis, Summary Measures, and Publication Bias
The process of our narrative synthesis review involved synthesizing the primary studies
by study features to aid the process of study descriptions and comparisons. Given our research
question, studies were synthesized according to the participants’ characteristics (i.e., grade),
reading outcomes, and duration of intervention, followed by quality appraisal.
For this synthesis, effect size estimates on reading related and reading achievement
outcomes were computed. Effect sizes were not aggregated because primary studies were
heterogeneous in terms of intervention content, application of interventions, and study quality.
Hence, rather than computing a pooled average effect size, an estimate of direction and size of
treatment effect was presented for each individual study in a tabular form, along with a summary
of the features of intervention studies reviewed. Effect sizes were calculated from the statistical
information in the original studies where sufficient statistical information was provided to
EFFECTS OF INDEPENDENT READING 12
calculate them. Hedges g, an effect size reflecting standardized mean differences between the
experimental and control groups on reading outcomes, was chosen because it provides a better
estimate of effect sizes than Cohen’s d on small sample sizes. Several studies in this review had
small sample sizes (i.e., < 50). The difference between the post-test mean of the treatment group
and the post-test mean of the control group was divided by the pooled standard deviation. A
positive Hedges g reflected intervention group outperforming control group. In addition to
presenting effect size estimates in a tabular form, a graphical presentation in a form of forest
plots was created to show estimates and their 95% confidence intervals for each study (but
without a pooled estimate) by participants’ characteristics (grade), reading outcomes, and
duration of intervention. Forest plots, created with a forestplot R package (version 1.10.1.;
Gordon et al., 2019), helped us visually explore relations within and between studies.
As to publication bias, we contacted authors for unreported outcomes. The author James
S. Kim was contacted via email (March 11, 2021) to obtain information about missing effect
sizes in studies Kim (2006), Kim and White (2008), and White and Kim (2008). The authors of
the present study have not received a response back.
Results
Study Characteristics
As indicated in Figure 1, a total of 14 experimental or quasi-experimental studies met the
selection criteria for inclusion in the synthesis. Overall, the studies included 5,522 participants in
the treatment group and 4,966 in the control group. A summary of the features of each
intervention study along with reading outcomes, measures, and reported or computed effect sizes
are listed in Table 1. Results are summarized and synthesized by participants’ characteristics
(grade), type of reading outcome, and duration of intervention (Table 1 and Figures 2-4),
EFFECTS OF INDEPENDENT READING 13
followed by a quality evaluation of the corpus of included studies. An illustration of Gersten et
al.’s (2005) and the WWC’s (2020) quality indicators in relation to the studies and tied to our
research question is provided in Figure 5.
Synthesis of Results
Effects by Participants’ Characteristics (Grade)
Ten studies were conducted in elementary grades (see Table 1). The main focus was 4th
and 5th grades. Four studies took place in middle or high school grades, with participants in 6th,
7th, 8th, and 10th grades. With the exception of three studies (Allington et al., 2010; Nunnery &
Ross, 2007; Nunnery et al., 2006), all studies conducted in elementary grades reported non-
significant positive effects of independent silent reading on reading achievement outcomes. In
middle school, the magnitude of effect sizes slightly increased compared to elementary school
although not consistently. In high school, the effects showed a picture of substantial increase,
with the majority of effects on reading achievement outcomes falling within the moderate range
(Cuevas et al., 2012; Cuevas et al., 2014). Figure 2 shows a forest plot depicting the direction
and magnitude of effect sizes aggregated by grade, with the majority of confidence intervals of
individual effect sizes in elementary school crossing the vertical line of no effect, and, as such,
indicating non-significant effects. On the other hand, the majority of confidence intervals of
effect sizes extracted in high school did not cross the line of no effect, signaling significant
effects.
Contrary to reading achievement outcomes, there was one reading outcome in which the
reported effects ranged from moderate to large already in elementary school. Kim (2007)
reported effects of g = 0.71 to 0.98 for the amount of reading across 1st through 3rd and also 5th
grade, but not 4th grade. Overall, results allude to an increasing trend in effects’ magnitude from
EFFECTS OF INDEPENDENT READING 14
elementary via middle to high school for reading achievement gains. Benefits of independent
reading practices on the amount of reading, however, might be observable already in lower and
upper elementary grades.
Effects by Type of Reading Outcome
Reading achievement. The majority of studies did not examine effects on any particular
reading outcome but instead focused on a general reading achievement outcome. Effect sizes
varied inconsistently across studies from non-significant to small, and moderate (see Table 1 and
Figure 3) regardless of the type of study design (experimental versus quasi-experimental).
Effects on reading achievement also varied substantially with different amount of time spent
reading. More time was not necessarily associated with stronger effects (e.g., Melton et al.,
2004). Lastly, use of standardized measures was not consistently associated with smaller effects,
as would be expected. One out of the two moderate effect sizes was detected with a measure of
generalized performance (Cuevas et al., 2014). Given the significant differences in the design of
the studies and duration of interventions, along with a disparate nature of reading achievement
assessments, it is difficult to generalize the findings and indicate to what extent independent
reading practice might be beneficial for reading achievement and under what circumstances.
Reading fluency, vocabulary, and reading comprehension. Of the studies that
evaluated effects of independent reading on specific reading outcomes, three experimental
studies examined the effects on reading fluency, one on vocabulary, and three on reading
comprehension (see Table 1 and Figure 3). Overall, student scores on reading fluency did not
significantly increase following the interventions. Effects ranged from g = -0.11. to g = 0.18 for
assessments designed to determine generalization effects on untrained text passages. Non-
significant effects of similar magnitude (range g = 0.03 – 0.25) were also found for vocabulary
EFFECTS OF INDEPENDENT READING 15
(Spichtig et al., 2019). However, it is difficult to evaluate what the effect sizes from the Spichtig
et al. (2019) study mean. The researchers described the comparison group as being engaged in
business-as-usual reading instruction during its 25-minute supplemental literacy block. The
control students were exposed to a variety of instructional approaches, including reading leveled
books in pairs or small groups, practicing silent reading, and receiving additional instruction in
word study, fluency, and reading strategies. Hence, potential intervention effects might have
been confounded. We cannot discern with a high degree of certainty that non-significant effects
were indeed a result of a non-effective intervention since the treatment group was not compared
against a control group in which students would read less and not be engaged in various forms of
reading practice in their classrooms.
Similarly to reading fluency, three experimental studies examined the impact of
independent reading on reading comprehension. In two studies, students made no significant
gains on comprehension, even though the overall time spent on independent reading was
substantial (e.g., over 3 summers; Allington et al., 2010). There was one study, however, that did
report significant moderate effect sizes on a standardized reading comprehension test (Cuevas et
al., 2012). Cuevas et al.’s (2012) findings indicated that 1 hour of weekly sustained silent reading
in high school provided substantial benefits in terms of stimulating reading comprehension.
However, also with this study, we need to be cautious when interpreting this effect. We cannot
parse out if reading comprehension questions administered after the intervention were a
component that led to such gains or if independent silent reading alone would materialize in
similar results. In other words, it is unclear whether the adjunct questions contributed to the
students’ gains beyond what students would have experienced by simply reading independently
for the hour.
EFFECTS OF INDEPENDENT READING 16
Amount of reading and print exposure. Aside from reading achievement outcomes,
four studies (see Table 1 and Figure 3) examined amount of reading and print exposure as
outcomes. In general, effect sizes were higher for those studies that measured amount of reading
done (Kim, 2007), compared to studies that used measures of print exposure, such as a title
recognition test (Pavonetti et al., 2003). It can be argued that interventions might have had
immediate effects on the amount of reading (e.g., Kim, 2007), however, impacts on the overall
long-term accumulated reading, as measured by title recognition test was less likely. It is difficult
to evaluate Pavonetti et al.’s (2003) results, though, because the study did not report monitoring
the amount of reading in the treatment as well as control groups. As such, it is possible that the
intervention just displaced reading that students would have done nonetheless. Overall, it appears
that the act of short-term independent reading alone will not result in students continuing to read
more in the long run, although this assumption has not been tested in any of the studies in the
current corpus.
Taken together, at face value, the evidence of benefit appears to be non-existent as most
effects detected were non-significant. Range of effect sizes aggregated by reading outcomes
(Figure 3) display that most confidence intervals cross the line of no effect, implying non-
significant gains from interventions. With many abovementioned limitations of the studies,
however, the results are difficult to evaluate, and no definite conclusions can be made. Further
investigation of studies implementing reliable, robust, replicable, and generalizable methods is
needed.
Effects by Duration of Intervention
As shown in Table 1 and Figure 4, the studies in this synthesis varied in the length of
intervention (range was from 8 weeks to 3 years). Five studies examined interventions with a
EFFECTS OF INDEPENDENT READING 17
duration of 8 weeks (2 months), with no reported information on the number of sessions per
week. Three studies implemented interventions over 18- to 24-week periods (4.5 – 6 months),
with approximately one session per week (Cuevas et al., 2012; Cuevas et al., 2014). Three
studies reported interventions which lasted from 32 weeks (8 months) up to 156 weeks (3 years),
again with no reported information on the intensity of sessions. Lastly, two studies did not
provide enough information to ascertain the duration of the intervention (Nunnery & Ross, 2007;
Pavonetti et al., 2003). As to the reported durations of individual sessions, only a few of the
studies provided this information (see Table 1). Typically, individual sessions of independent
reading lasted between 15 to 60 min.
Interventions with a duration of up to 8 weeks and over 32 weeks yielded non-significant
effects for reading outcomes. Interventions of 18 to 24 weeks reported mixed results, with the
effects ranging from non-significant (Shannon et al., 2015) to small (Shannon et al., 2015;
Siddiqui et al., 2016), and moderate (Cuevas et al., 2012; Cuevas et al., 2014). The differences in
effects may be explained by several factors, including outcome measures employed. Two
moderate effects were detected in Cuevas et al. (2012) on researcher-developed reading
assessments. Experimenter- or program-developed measures inherent to treatment and/or
measures closely aligned with the intervention, such as the ones used in Cuevas et al. (2012),
have been found to be associated with much larger effect sizes than measures that are
independent of treatments (Slavin & Madden, 2011).
All in all, many studies failed to report the length, intensity, and frequency of
intervention sessions. Of the studies that did report them, the effects ranged from non-significant
to small, and moderate in terms of educational importance (see also Figure 4 for variability in
effect sizes). More generally, however, it is difficult to evaluate the value of these gains without
EFFECTS OF INDEPENDENT READING 18
more information about the length and frequency of the programs. Moreover, caution is
warranted on the meaning of these results because studies (e.g., Cuevas et al., 2012; Cuevas et
al., 2014; Siddiqui et al., 2016) failed to provide information about the time spent during the
interventions by the control group and descriptions of how that time was used.
Assessment of Internal Validity of Individual Studies
We used Gersten et al.’s (2005) and WWC (2020) quality indicators to assess the
methodological quality of studies. Figure 5 shows a quality evaluation of primary studies using a
“traffic light” plot, which tabulates the judgement for each individual study in each quality
appraisal domain. The green color indicates a study aligned well with a quality appraisal domain,
whereas the red color means a study adhered poorly to quality indicators.
Examining each individual quality appraisal domain, the quality indicator for describing
participants was aimed at assessing maintenance of experimental control. To be able to establish
experimental causality, students in both conditions should be equivalent on reading engagement
and amount of independent reading before the intervention. As shown in Figure 5, none of the
reviewed studies reported data on the baseline equivalence of the amount of reading. The closest
approximation to this quality indicator was found in three studies (21%). In these studies, the
authors reported on the attitude towards reading before the intervention which could be
considered a proxy for prior reading engagement and amount of reading.
As to quality indicators for implementation of the intervention and description of
comparison conditions, Figure 5 shows that ten studies (71%) implemented interventions that
were clearly described, five studies (36%) described and assessed implementation fidelity, eight
studies (57%) reported appropriate levels of overall attrition and no differential attrition between
groups, three studies (21%) described services provided in the control group and reported
EFFECTS OF INDEPENDENT READING 19
keeping track of the amount of reading that students in both groups did during the interventions,
and five studies (36%) reported providing no additional independent reading instruction in
control groups. Given low percentages of the number of studies adhering to the quality indicator
for implementation of the intervention and description of comparison conditions, we can
conclude that the validity of causal inference was threatened in majority of studies.
Lastly, regarding quality indicators for outcomes measures, 12 studies (86%) used more
than one assessment or an assessment of generalized performance to measure a reading
achievement outcome. However, the reporting of assessments’ reliability (eight studies; 57%)
and validity estimates (three studies; 21%) was not highly aligned with quality indicators, which
invariably weakens the power of studies.
The quality evaluation shows many weaknesses and limitations that studies in our corpus
display. Three perhaps most essential indicators relevant to our research question showed poor
quality alignment. First, 21% of studies met baseline equivalence on the amount of reading,
second, 21% of studies described the nature of services provided in comparison conditions, and,
third, 36% of studies reported control conditions which would not confound causal conclusions.
Based on the number of indicators that aligned with quality standards as evaluated in the present
report, studies by Kim (2007), Kim and White (2008), and White and Kim (2008) provided the
clearest investigations examining potential effects of independent reading on reading outcomes.
These authors reported non-significant impact of such practices. On the whole, however, the
studies were not likely to meet quality criteria. As such, it appears that even 20 years after the
NRP report (2000), there is still insufficient and incomplete rigorous evidence on the benefit of
independent reading on reading outcomes. Until such evidence is forthcoming, we cannot
indicate that research has proven or unproven that such procedures actually work.
EFFECTS OF INDEPENDENT READING 20
Publication Bias and Adverse and Harmful Effects
Theses, dissertations, evaluation reports, conference presentations and submissions, and
any unpublished manuscripts were excluded from this narrative synthesis review. We elected to
select only studies that have undergone the peer-review process to ensure the findings aligned
with evidence-based practices. This decision might have introduced bias in our report. For
interested readers, however, we still provide a summary and quality appraisal of intervention
studies published as theses and dissertations in Supplemental Materials. No adverse or harmful
effects were identified in the corpus of primary studies.
Discussion
Silent independent reading is a commonly applied practice in schools. Efforts to uncover
its effectiveness have been underway for at least 30 years. This endeavor culminated with the
release of the NRP’s synthesis report in 2000, which yielded initial, yet inconclusive findings,
mostly due to poor research design being implemented by the primary studies. The present study
provides an update and extension of the reading fluency subsection of the NRP’s (2000) report
by conducting a narrative synthesis review and quality evaluation of the corpus of literature
published between 2000 and 2020. At face value, the evidence for the effectiveness of silent
independent reading on reading outcomes appears to be non-existent. However, due to a lack of
primary studies adhering to the highest quality standards and implementation, it is impossible to
determine whether such a result is universal or whether there might be conditions under which
silent independent reading could have meaningful beneficial effects.
Reported effect sizes provided some evidence to allude to an increasing effect of
independent reading on reading outcomes by grade. This finding is in line with the results from
the NRP’s fluency synthesis review on independent reading (2000) that found that benefits of
EFFECTS OF INDEPENDENT READING 21
independent reading were strongest among older students. An increasing effect by grade is not a
surprising finding. Reading theories (e.g., Perfetti, 1985) posit that decoding skills, typically
acquired in elementary school, must first be effective before children can use wide reading as a
tool to attain new knowledge in content-area classes in middle and high school.
A picture of increasing effect by grade might be directly related to reading outcomes
being examined. Our results revealed that if there is evidence of a potential effect of voluntary
reading on reading outcomes, it will likely be reflected in gains in reading comprehension and
general reading achievement (i.e., skills important for content-area reading in middle and high
school), rather than code-related skills, such as reading fluency (i.e., skills learned in elementary
school). Previous longitudinal research corroborates these findings to show that gains in reading
comprehension were incidental to voluntary reading practices, whereas improvements in reading
fluency were not (Torppa et al., 2019). It is important to keep in mind, though, that primary
studies which showed small to moderate beneficial effects on reading achievement outcomes
revealed methodological limitations. As such, we need to take care not to overreach in the
interpretation of these findings. In contrast to reading achievement outcomes, more pronounced
benefits of independent reading were evident in increases in the amount of reading. One of the
methodologically strongest designed studies from our corpus of studies (Kim, 2007) found
significant gains in amount of reading in as little as 8 weeks of intervention. At first glance, this
result is encouraging. However, the impact appears to be short-lived as evidenced by non-
significant effects on a title recognition test, a proxy for long-term accumulated reading.
Based on our data, it was difficult to establish optimal instructional time and/or amount
of (daily) independent reading needed to promote gains in reading outcomes. Even
methodologically well-designed studies failed, on average, to report duration, frequency, and
EFFECTS OF INDEPENDENT READING 22
intensity of intervention and monitor the amount of reading in both groups. Likely, there is an
upper limit on the amount of time spent reading that will continue to be effective in building
reading skills. As our primary studies indicated (e.g., Allington et al., 2010; Melton et al., 2004),
simply increasing time did not automatically lead to greater reading achievement. In an analysis
of reading growth from second to fifth grade, Anderson (1988) found that the relation between
time spent reading books and reading achievement was not linear. Reading comprehension rose
sharply with up to about 10 minutes of daily reading and then leveled off. Overall, our data
coupled with Anderson’s (1988) observation suggest that the largest gains in reading are likely
not a result of considerable time devoted to voluntary reading, but other factors that have been
strongly linked to reading behaviors, such as general reading performance (e.g., Erbeli et al.,
2020).
Although the present synthesis review holds promise in helping move research into
utilizing high quality experimental designs measuring the effect of silent independent reading,
the conclusions that can be drawn from it are limited due to the quality of primary studies. Per
Gersten et al. (2005) and WWC’s (2020) guidelines, three quality indicators, which were key to
our research question and were also identified as such by the NRP panel, were met infrequently:
1. quality indicator for describing participants – baseline equivalence of the amount of reading,
2. quality indicator for implementation of the intervention and description of comparison
conditions – description of services provided in comparison conditions, 3. quality indicator for
implementation of the intervention and description of comparison conditions – documentation of
the nature of instruction provided in comparison group. As to the first indicator, with the
exception of three studies (Kim, 2007; Kim & White, 2008; White & Kim, 2008), studies did not
take the baseline equivalence into account. As such, studies could not causally demonstrate
EFFECTS OF INDEPENDENT READING 23
whether any increases in the amount of reading resulted from the intervention or stemmed from
the existence of uncontrolled extraneous variables. A scenario of not controlling for an
extraneous variable would threaten the validity of the interventions. Regarding the second
indicator, more high-quality randomized control studies are needed which would monitor the
amount of reading in both groups. Else, a similar scenario as with the first indicator is likely to
unfold. As to the third indicator, with the exception of three studies (Allington et al., 2010; Kim,
2006; Kim, 2007), additional independent reading instruction was provided in the control groups.
As such, if routines of independent reading are not compared against procedures in which
students actually read less, then implications of independent reading cannot be validly
established. Ruling out any instructional effects as potential confounders of the effect of an
exposure to independent reading would provide an instance of a more rigorous experimental
design.
Our study contributes some imperative educational implications. Results of the NRP
reading fluency synthesis review on independent reading (2000) and our synthesis show that at
this point in time it is impossible to ultimately establish that silent independent reading does not
work under any conditions. It appears that independent reading alone might not be a viable way
for improved reading outcomes. Yet, prior research suggests that sustained silent reading in
conjunction with engagement, scaffolding, and elements of social supports might be effective
(Cuevas et al., 2012; Kim, 2006; Kim & White, 2008). Moreover, for students with reading
disabilities, research shows that requiring more intensive interventions, extensive reading,
including direct and explicit reading instruction, as well as engaged reading practices with
effective feedback might lead to reading advancement (Wexler et al., 2010). Future work has a
huge task on itself to investigate the conditions under which wide reading practices may matter.
EFFECTS OF INDEPENDENT READING 24
Once we have rigorous research in place and identify the details under which independent
reading practices might lead to the biggest gains for most students, the next step would be to
evaluate their effectiveness in a longitudinal experimental framework on different levels of
students.
Conclusion
The present study adds to an existing body of research examining the impact of silent
independent reading on reading outcomes. Given the inconclusive findings, our study might open
new lines of investigation underlying the potential effects of free time reading. Even though the
evidence is incomplete, silent independent reading might still have clear value for improving
reading outcomes, as it could work under specific conditions. The onus is on researchers to
identify these conditions by implementing methodologically rigorous randomized control trials,
which will provide avenues for discovery of benefits of voluntary reading on reading outcomes.
EFFECTS OF INDEPENDENT READING 25
References
References marked with an asterisk indicate studies included in the synthesis review.
*Allington, R. L., McGill-Franzen, A., Camilli, G., Williams, L., Graff, J., Zeig, J., Zmach, C.,
& Nowak, R. (2010). Addressing summer reading setback among economically
disadvantaged elementary students. Reading Psychology, 31(5), 411-427
Anderson, R. C., Wilson, P. T., & Fielding, L. G. (1988). Growth in reading and how children
spend their time outside of school. Reading Research Quarterly, 23(3), 285–303.
*Cuevas, J. A., Irving, M. A., & Russell, L. R. (2014). Applied cognition: Testing the effects of
independent silent reading on secondary students’ achievement and attribution. Reading
Psychology, 35(2), 127-159.
*Cuevas, J. A., Russell, R. L., & Irving, M. A. (2012). An examination of the effect of
customized reading modules on diverse secondary students’ reading comprehension and
motivation. Educational Technology Research and Development, 60(3), 445-467.
Erbeli, F., van Bergen, E., & Hart, S. A. (2020). Unraveling the relation between reading
comprehension and print exposure. Child Development, 91, 1548-1562.
Gersten, R., Fuchs, L. S., Compton, D., Coyne, M., Greenwood, C., & Innocenti, M. S. (2005).
Quality indicators for group experimental and quasi-experimental research in special
education. Exceptional Children, 71, 149–164. doi:10.1177/001440290507100202
Gordon, M., Lumley, T., Gordon, M.M. (2019). Package forestplot (Version 1.10.1) [Computer
software]. CRAN. https://cran.r-project.org/web/packages/forestplot/forestplot.pdf
Hiebert, E., & Reutzel, D. R. (2014). Revisiting silent reading: New directions for teachers and
researchers. International Reading Association.
EFFECTS OF INDEPENDENT READING 26
*Kim, J. S. (2006). Effects of a voluntary summer reading intervention on reading achievement:
Results from a randomized field trial. Educational Evaluation and Policy Analysis, 28(4),
335-355.
*Kim, J. S. (2007). The effects of a voluntary summer reading intervention on reading activities
and reading achievement. Journal of Educational Psychology, 99(3), 505-515.
*Kim, J. S., & White, T. G. (2008). Scaffolding voluntary summer reading for children in grades
3 to 5: An experimental study. Scientific Studies of Reading, 12(1), 1-23.
LaBerge, D., & Samuels, S. J. (1974). Toward a theory of automatic information processing in
reading. Cognitive Psychology, 6(2), 293-323.
McGuinness, L. A., & Higgins, J. P. (2021). Risk‐of‐bias Visualization (robvis): An R package
and shiny web app for visualizing risk‐of‐bias assessments. Research Synthesis Methods,
12(1), 55-61.
*Melton, C. M., Smothers, B. C., Anderson, E., Fulton, R., Replogle, W. H., & Thomas, L.
(2004). A study of the effects of the Accelerated Reader program on fifth grade students'
reading achievement growth. Reading Improvement, 41(1), 18-24.
Merga, M. K. (2013). Should silent reading feature in a secondary school English programme?
West Australian students’ perspectives on silent reading. English in Education, 47, 229.
National Reading Panel. (2000). Teaching children to read: An evidence-based assessment of the
scientific research literature on reading and its implications for reading instruction.
National Institute of Child Health and Human Development.
*Nunnery, J. A., & Ross, S. M. (2007). The effects of the School Renaissance program on
student achievement in reading and mathematics. Research in the Schools, 14(1), 40-59.
EFFECTS OF INDEPENDENT READING 27
*Nunnery, J. A., Ross, S. M., & McDonald, A. (2006). A randomized experimental evaluation of
the impact of Accelerated Reader/Reading Renaissance implementation on reading
achievement in grades 3 to 6. Journal of Education for Students Placed at Risk, 11(1), 1-
18.
*Pavonetti, L. M., Brimmer, K. M., & Cipielewski, J. F. (2002). Accelerated Reader: What are
the lasting effects on the reading habits of middle school students exposed to Accelerated
Reader in elementary grades? Journal of Adolescent & Adult Literacy, 46(4), 300-311.
Perfetti, C. A. (1985). Reading ability. Oxford University Press.
*Shannon, L. C., Styers, M. K., Wilkerson, S. B., & Peery, E. (2015). Computer-assisted
learning in elementary reading: A randomized control trial. Computers in the
Schools, 32(1), 20-34.
*Siddiqui, N., Gorard, S., & See, B. H. (2016). Accelerated Reader as a literacy catch-up
intervention during primary to secondary school transition phase. Educational
Review, 68(2), 139-154.
Slavin, R., & Madden, N. A. (2011). Measures inherent to treatments in program effectiveness
reviews. Journal of Research on Educational Effectiveness, 4(4), 370-380.
*Spichtig, A. N., Gehsmann, K. M., Pascoe, J. P., & Ferrara, J. D. (2019). The impact of
adaptive, web-based, scaffolded silent reading instruction on the reading achievement of
students in grades 4 and 5. The Elementary School Journal, 119(3), 443-467.
Torppa, M., Niemi, P., Vasalampi, K., Lerkkanen, M. K., Tolvanen, A., & Poikkeus, A. M.
(2020). Leisure reading (but not any kind) and reading comprehension support each
other—A longitudinal study across grades 1 and 9. Child Development, 91(3), 876-900.
EFFECTS OF INDEPENDENT READING 28
Vaughn, S., Elbaum, B. E., Wanzek, J., Scammacca, N., & Walker, M. A. (2014). Code sheet
and guide for education-related intervention study syntheses. The Meadows Center for
Preventing Educational Risk.
Wexler, J., Vaughn, S., Roberts, G., & Denton, C. A. (2010). The efficacy of repeated reading
and wide reading practice for high school students with severe reading
disabilities. Learning Disabilities Research & Practice, 25(1), 2-10.
What Works Clearinghouse. (2010). Accelerated ReaderTM. Adolescent literacy intervention
report. U.S. Department of Education.
What Works Clearinghouse. (2020). Procedures and standards handbook (Version 4.1). Institute
of Education Sciences.
*White, T. G., & Kim, J. S. (2008). Teacher and parent scaffolding of voluntary summer
reading. The Reading Teacher, 62(2), 116-125.
EFFECTS OF INDEPENDENT READING 29
Table 1
Summary of the Features of Intervention Studies Reviewed and Their Findings with Effect Sizes
Study
Design
Grade
level(s)
Readi
ng
Outco
me(s)
Measure(s)
N T, N C for
the outcome
measure(s)
Intervention
duration
Intervention
frequency (length
of each session in
min: total number
of sessions)
Instruc
tional
time
Reading
Program
Effect Size
Allington
et al., 2010
RCT
3, 4, 5
RC
FCAT
852, 478
3 years5
NR: NR
SU
N/A
g = 0.14*
Cuevas et
al., 2012
RCT
10
RC
GM (RC a covariate in
the model)
38, 45
18 weeks
60min: 14
sessions
SY
SSR
(ISR)
g = 0.58*
10
RC
GM (RA a covariate in
the model)
38, 43
g = 0.55*
10
RC1
Researcher developed
RC assessment
31, 56
g = 0.23
10
RC2
Researcher developed
RC assessment
31, 56
g = 0.24
10
RC3
Researcher developed
RC assessment
31, 56
g = 0.33
10
RC4
Researcher developed
RC assessment
31, 56
g = 0.50*13
10
RM
AMRS
43, 51
g = 0.59*13
Cuevas et
al., 20141
RCT
10
RA
GETRAL
39, 44
18 weeks
60min: 14
sessions
SY
SSR
(ISR)
g = 0.49*
Kim, 2006
RCT
4
RA
ITBS
252, 234
8 weeks
NR: NR
SU
N/A
g = 0.18
4
RF
DIBELS ORF
NR, NR
Not sufficient information
to compute g
4
AmR
LHS
252, 234
g = 0.02
Kim, 2007
RCT
1, 2, 3, 4, 5
RA
SAT-10, entire sample
141, 138
8 weeks
NR: NR
SU
N/A
g = 0.04
1
RA
SAT-10, 1st grade
NR, NR
g = -0.01
2
RA
SAT-10, 2nd grade
NR, NR
g = 0.07
3
RA
SAT-10, 3rd grade
NR, NR
g = 0.13
4
RA
SAT-10, 4th grade
NR, NR
g = -0.34
5
RA
SAT-10, 5th grade
NR, NR
g = 0.31
1, 2, 3, 4,
5
AmR
LHS, entire sample
137, 139
g = 0.71*
1
AmR
LHS, 1st grade
32, 28
g = 0.98*
2
AmR
LHS, 2nd grade
23, 22
g = 0.75*
3
AmR
LHS, 3rd grade
30, 34
g = 0.78*
4
AmR
LHS, 4th grade
19, 22
g = -0.02
5
AmR
LHS, 5th grade
33, 33
g = 0.89*
EFFECTS OF INDEPENDENT READING 30
Table 1
cont.
Study
Design
Grade
level(s)
Readi
ng
Outco
me(s)
Measure(s)
N T, N C for
the outcome
measure(s)
Intervention
duration
Intervention
frequency (length
of each session in
min: total number
of sessions)
Instruc
tional
time
Reading
Program
Effect Size
Kim &
White,
2008
RCT
3, 4, 5
RA
ITBS
93, 107
1 lesson at
school +8
weeks
45min at school +
NR: NR
SY +
SU
N/A
g = 0.02
3, 4, 5
RF
DIBELS ORF
81, 91
g = -0.11
3, 4, 5
AmR
LHS
NR, NR
Not sufficient information
to compute g; as reported,
the effect was not
significant
Melton et
al., 2004
QED
5
RA
TNSRAT, quartile 110
(adjusted means)
55, 75
1 year2
NR: NR
SY
AR
g = 0.31
5
RA
TNSRAT, quartile 2
(adjusted means)
93, 66
g = -0.78*
5
RA
TNSRAT, quartile 3
(adjusted means)
104, 66
g = -0.62*
5
RA
TNSRAT, quartile 4
(adjusted means)
70, 63
g = -1.03*
Nunnery &
Ross, 2007
QED
5
RA
TAAS, 5th grade,
cohort 1
450, 462
NR
NR: NR
SY
AR
g = 0.21*
5
RA
TAAS, 5th grade,
cohort 2
442, 449
g = 0.14*
8
RA
TAAS, 8th grade,
cohort 1
400, 448
g = 0.00
8
RA
TAAS, 8th grade,
cohort 2
482, 510
g = 0.02
Nunnery et
al., 2006
RCT
3
RA
STAR, 3rd grade
138, 112
8 months8
30 to 60 min: NR
SY
AR
g = 0.33*12
4
RA
STAR, 4th grade
176, 176
g = 0.1312
5
RA
STAR, 5th grade
121, 94
g = 0.0512
6
RA
STAR, 6th grade
73, 59
g = 0.0312
Pavonetti et
al., 2003
QED
7
PR
TRT, entire sample,
adjusted scores
836, 700
NR
NR: NR
SY
AR
g = -0.01
7
PR
TRT, district 1
114, 183
g = -0.19
7
PR
TRT, district 2
364, 138
g = -0.39*
7
PR
TRT, district 3
358, 379
g = 0.25*
EFFECTS OF INDEPENDENT READING 31
Table 1
cont.
Study
Design
Grade
level(s)
Readi
ng
Outco
me(s)
Measure(s)
N T, N C for
the outcome
measure(s)
Intervention
duration
Intervention
frequency (length
of each session in
min: total number
of sessions)
Instruc
tional
time
Reading
Program
Effect Size
Shannon et
al., 2015
RCT
1, 2, 3, 4
RA
STAR, entire sample
181, 145
24 weeks
37min9: NR
SY
AR
g = 0.2011
1
RA
STAR, 1st grade
57, 25
g = 0.0411
2
RA
STAR, 2nd grade
36, 33
g = 0.4511
3
RA
STAR, 3rd grade
51, 31
g = 0.2111
4
RA
STAR, 4th grade
37, 56
g = 0.1811
Siddiqui et
al., 2016
RCT
63
RA
NGRT
175, 164
20 weeks
NR: NR
SY
AR
g = 0.24*
Spichtig et
al., 2019
RCT
4, 5
RA
GRADE, entire
sample
213, 213
25 hours6
15.4/157 min:
87/99 sessions
SY
Reading
Plus
g = 0.14
4
RA
GRADE, 4th grade
105, 105
g = 0.08
5
RA
GRADE, 5th grade
108, 108
g = 0.23
4, 5
RC
GRADE, entire
sample
213, 213
g = 0.14
4
RC
GRADE, 4th grade
105, 105
g = 0.11
5
RC
GRADE, 5th grade
108, 108
g = 0.18
4, 5
Voc
GRADE, entire
sample
213, 213
g = 0.13
4
Voc
GRADE, 4th grade
105, 105
g = 0.03
5
Voc
GRADE, 5th grade
108, 108
g = 0.25
4, 5
RF
VG, entire sample
165, 165
g = 0.15
4
RF
VG, 4th grade
75, 75
g = 0.17
5
RF
VG, 5th grade
90, 90
g = 0.18
White &
Kim, 2008
RCT
3, 4, 5
RA
ITBS
93, 107
1 lesson at
school +8
weeks
45min at school +
NR: NR
SY +
SU
N/A
g = 0.02
3, 4, 5
RF
DIBELS ORF
81, 91
g = -0.11
3, 4, 5
AmR
LHS
NR, NR
Not sufficient information
to compute g; as reported,
the effect was not
significant
Note. * = p < .05, NR = not reported, RCT = experimental design, QED = quasi-experimental design, N/A = not applicable, N = number of individual students; T = treatment
group, C = control group, RC = reading comprehension, RM = reading motivation, RA = reading achievement, RF = reading fluency, PR = print exposure, AmR = amount of
EFFECTS OF INDEPENDENT READING 32
reading, Voc = vocabulary, FCAT = Florida Comprehensive Achievement Test, GM = Gates-MacGinitie, AMRS = Adult Motivation for Reading Survey, GETRAL = Georgia
End-Of-Course Test for Reading and American Literature, ITBS = Iowa Test of Basic Skills, DIBELS ORF = Dynamic Indicators of Basic Early Literacy Skills Oral Reading
Fluency, LHS = Literacy Habits Survey, SAT-10 = Stanford Achievement Test, 10th ed, TNSRAT = Terra Nova Standardized Reading Achievement Test, TAAS = Texas
Assessment of Academic Skills, STAR = STAR Reading Test from the program Accelerated Reader, TRT = Title Recognition Test, NGRT = New Group Reading Test, GRADE
= Group Reading Assessment Diagnostic Evaluation, VG = Visagraph, SU = summer, SY = school year, SSR (ISR) = Sustained Silent Reading (Independent Silent Reading), AR
= Accelerated Reader.
1 = Cuevas et al., 2014 is an identical study to Cuevas et al., 2012 (i.e., same participants, same intervention). The only difference between the two studies is the use of different
reading outcome measures.
2 = The study reports one year of participation in the reading program Accelerated Reader. However, the exact number of weeks or months over the year (52 weeks including
summer or 40 weeks over a regular school year) is not reported.
3 = The authors conducted the study in year 7 in the UK, which corresponds to grade 6 in the US.
4 = White & Kim, 2008 presents a summary of results as reported in Kim & White, 2008. As such, the results reported here are identical across both Kim &White, 2008 and White
& Kim, 2008 studies.
5 = The study reports that the intervention lasted for 3 years. It reports providing books to the treatment group for 3 consecutive summers. However, the exact number of weeks or
months is not reported.
6 = The exact number of weeks is not reported. The study reports that the intervention lasted 25 hours during the school year.
7 = The first two numbers refer to the averages for minutes per session in fourth and fifth grade interventions, respectively. The next two numbers refer to the total number of
sessions in fourth and fifth grade interventions, respectively.
8 = The exact number of weeks is not reported. The study reports that the intervention lasted from September to April.
9 = 37 min refers to the average length of the sessions. The range reported was 30-45 min.
10 = Quartiles refer to reading comprehension performance on the Terra Nova, with quartile 1 representing the lowest and quartile 4 the highest score.
11 = The authors reported a moderate effect size of d = 0.38. We could not locate information in the article to recalculate and replicate that effect size. Instead, we report the
adjusted effect sizes across grades using information from descriptive analysis.
12 = The authors reported effect size estimates of 0.36, 0.16, 0.09, 0.09 for grades 3 through 6. They reported these results paralleled those of the descriptive analysis. We
recalculated the effect sizes using information from descriptive analysis. Here we report the adjusted effect sizes.
13 = The authors reported non-significant effects. We could not locate information in the article to replicate the non-significance. Here, we report the calculated effect sizes using
information from descriptive analysis.
EFFECTS OF INDEPENDENT READING 33
Figure 1
Flowchart of the Search and Screening Process
EFFECTS OF INDEPENDENT READING 34
Figure 2
Forest Plot of Standardized Mean Differences on Reading Outcomes Between Experimental and Control Groups Aggregated by Participants’
Characteristics (Grade)
Note. Studies for which effect sizes could not be computed due to missing data are omitted from the figure.
EFFECTS OF INDEPENDENT READING 35
Figure 3
Forest Plot of Standardized Mean Differences on Reading Outcomes Between Experimental and Control Groups Aggregated by Reading Outcomes
Note. Studies for which effect sizes could not be computed due to missing data are omitted from the figure. RC = reading comprehension, RM = reading motivation, RA = reading
achievement, RF = reading fluency, PR = print exposure, AmR = amount of reading, Voc = vocabulary, FCAT = Florida Comprehensive Achievement Test, GM = Gates-
MacGinitie, AMRS = Adult Motivation for Reading Survey, GETRAL = Georgia End-Of-Course Test for Reading and American Literature, ITBS = Iowa Test of Basic Skills,
DIBELS ORF = Dynamic Indicators of Basic Early Literacy Skills Oral Reading Fluency, LHS = Literacy Habits Survey, SAT-10 = Stanford Achievement Test, 10th ed, TNSRAT
= Terra Nova Standardized Reading Achievement Test, TAAS = Texas Assessment of Academic Skills, STAR = STAR Reading Test from the program Accelerated Reader, TRT
= Title Recognition Test, NGRT = New Group Reading Test, GRADE = Group Reading Assessment Diagnostic Evaluation, VG = Visagraph, SU = summer, SY = school year,
SSR (ISR) = Sustained Silent Reading (Independent Silent Reading), AR = Accelerated Reader.
EFFECTS OF INDEPENDENT READING 36
Figure 4
Forest Plot of Standardized Mean Differences on Reading Outcomes Between Experimental and Control Groups Aggregated by Intervention
Duration
Note. Studies for which effect sizes could not be computed due to missing data are omitted from the figure.
EFFECTS OF INDEPENDENT READING 37
Figure 5
Gersten et al. (2005) and WWC (2020) Quality Indicators by Primary Studies
EFFECTS OF INDEPENDENT READING 38
Note. Cuevas et al., 2012 = Information on internal consistency (alpha = .86) was reported for one measure, The Adult Motivation for Reading Survey. No other information on
reliability and validity was reported. Kim, 2006 = Information on validity was reported for the Literacy Habits Survey, but not for the Iowa Test of Basic Skills or The Dynamic
Indicators of Basic Early Literacy Skills Oral Reading Fluency. Allington et al. (2010) and Kim (2006) = The authors note a completion of the Literacy Habits Survey, which
serves as a test to monitor the amount of reading in both conditions. However, limited information is reported from it to draw any conclusions about the nature of services provided
in the comparison conditions. Kim, 2007 and Kim & White (2008) = Baseline equivalence of the amount of independent reading was not explicitly reported. However, the authors
tested for attitude towards reading before the intervention, using The Elementary Reading Attitude Survey (ERAS), which could be considered a proxy for reading engagement in
reading. There were no significant differences on the attitude towards reading between the treatment and control groups. The ERAS was also administered in Kim (2006), however
with limited data from it reported to draw any conclusions. Spichtig et al. (2019) = Information on alternate form reliability (r = .87 to .89) was reported for one measure, The
Group Reading Assessment Diagnostic Evaluation. No other information on reliability and validity were reported.
