Approaches to Parent Involvement for Improving the Academic Performance of Elementary School Age Children

Findings from Primary Studies

The role that parents play in the academic achievement of their children has long been thought to be a centrally important one. However, it was not until the 1960's that the effects of programs designed to foster the role of parent involvement were systematically studied using an experimental design. Evaluation of the Head Start Program in the United States focused nationally on outcomes related to parent involvement (Coleman, Campbell, Hobson, McPartland, Mod, Weinfeld, & York, 1966). The results of this evaluation suggested a substantial relationship between the parent's involvement in their child's education and their child's success in academic domains. Subsequent studies supported the findings from Coleman, et al. (Duff & Adams, 1981; Henderson, 1987; 1988). Even so, quantitative evidence on the effect of parent involvement on student achievement has been mixed. For instance, researchers have reported effect sizes ranging from positive to negative to no significant differences between experimental and control groups (Griffith, 1996; Heller, & Fantuzzo, 1993; Henry, 1974; Keith, Reimers, Ferman, Pottenbaum, & Aubrey, 1986; Ryan, 1964; Searles, Lewis & Morrow, 1982).

Some of the discrepancies across studies can be explained by the nature of the data collection and research design. For example, some investigators studied the relationship between parent involvement and a child's school success using direct observation (Arbuckle & MacKinnon, 1988), surveys or questionnaires, or both (Edwards & Warin, 1999). Other investigators mounted traditional experimental designs to compare student performance across randomly allocated groups (DeBaryshe, 1993; Woods, Barnard, & TeSelle, 1974).

Another potential source of discrepancy in the findings across studies relates to outcomes measured. Specifically, a variety of dependent variables have been reported in studies on parent involvement including reading achievement (Epstein, 1987; 1991; Tizard, Schofield, & Hewison, 1982; Trovato & Bucher, 1980; Walberg, Bole, & Waxman, 1980; Woods, Barnard, & TeSelle, 1974), math achievement (Fantuzzo, Davis, & Ginsburg, 1995; Heller & Fantuzzo, 1993; Morgan & Sorensen, 1999), and perceptual skills training (Garrison, 1977). Furthermore, differences in study findings and the conclusions often do not take into account other important factors that can affect the validity of study findings such as the reliability of scales and tests; controlling for important child background characteristics such as grade, age, socioeconomic status (SES); and controlling for important parent background characteristics such as SES, education and training (Reynolds, Weissberg, & Kasprow, 1992).

In addition, researchers have defined parent involvement differently or so broadly that it is difficult to understand how to measure it consistently. For example, one group of researchers defined parent involvement as parent participation in educational activities at both school and home (Christenson, Rounds, & Gorney, 1992). Epstein (1987) suggested that parent involvement is multi-dimensional and included: (1) parents providing a home environment that supports learning, (2) communication between parents and teachers on classroom performance, (3) parent's active attendance at school activities such as PTA, (4) parent's engagement and monitoring of home learning activities, and (5) parent's participation in school-based decision making such as school committees.

Typically, the study of parent involvement in home-school environments has identified educational activities that ranged from parent participation in parent-teacher organizations and conferences to specific training in activities designed to enhance an academic skill such as reading or math proficiency. The vast majority of the studies on parent involvement have focused on elementary school age children (Keith & Cool, 1992). Most schools in the US and abroad have some component of parent involvement integrated into a child's daily activities. Thus, differentiating the effects of parent involvement from other factors that affect student achievement can be difficult.

Findings from Narrative Reviews

In addition to the many primary studies on the statistical relationship between parent involvement and student achievement, several narrative summaries of parent involvement have been published over the past 20 years. Gordon (1977), for example, reviewed parent involvement programs and proposed a three dimensional model: (1) Parent Impact, (2) School Impact, and (3) Community Impact. Gordon's summary indicated that all three models produced positive effects, however, the summary was made without a quantitative analysis of the degree or nature of the intervention effects for any model. This narrative summary provided promising framework to further examine the effect of parent involvement on their children's academic performance.

In a more comprehensive review than Gordon's (1977), Christenson, Rounds, and Gorney (1992) examined over 160 studies that described variations on parent involvement. These reports focused on family components of parent expectations and attributions, structure for learning, home affective environment, discipline, and parent involvement. Christenson, et al. concluded that while there is a positive correlation between student achievement and parent components, there are background variables underlying parent involvement components and student achievement. That is to say, the influence of other variables such as SES, ethnicity, or gender had not been adequately accounted for while studying the interactive nature of the intervention of parent involvement and student achievement.

In a recent synthesis, Mattingly, Prislin, McKenzie, Rodriguez, and Kayzar (2002) reviewed a total of 41 studies described as ‘evaluations’ of parent involvement programs. While this review did not provide a meta-analytic summary of these studies, the authors concluded that there was little evidence to support the efficacy of parent involvement programs to improve student achievement. Mattingly, et al. indicate that their results do not suggest that parent involvement programs are not effective but that the quality of the evidence is compromised due to methodological weaknesses.

Findings from Meta-Analyses

The first meta-analysis on the topic of parent involvement was conducted by Graue, Weinstein, and Walberg (1983) who quantitatively synthesized 29 empirical studies with various types of study designs conducted from 1970-1980 that focused on parent involvement and student academic success. Graue, et al. assessed a wide variety of parent involvement studies without specifically measuring the effect of a particular type of parent involvement, such as joint book reading or homework. Graue, et al. concluded that there was a significant academic advantage for those children whose parents participated in their child's educational program. However, the researchers did not investigate study design type (e.g., experimental versus quasi-experimental) or intervening variables (e.g., length of parent involvement).

Several meta-analyses have been reported in the area of parent involvement and can generally be categorized into one of two types of studies: correlational meta-analyses that assess the nature of the relationship between parent involvement activity and measured outcomes; or intervention effect meta-analyses that assess the impact of a formal program to change measured outcomes. We recognize that several correlational meta-analyses have contributed much to our understanding of the nature of the contributing variables and conditions. Rosenzweig (2000) assessed the association between parent practices and school success across 34 studies representing seven classes of outcome variables. She concluded that the 20 specific parent practices could account for 23% of the explained variance in children's academic performance. No attempt was made to address the causal impact of parent involvement on student achievement.

Similarly, Fan and Chen (2001) reviewed 25 studies that focused on the relationship between parent involvement and academic performance. They found that the overall strength of relationship was modest r=.25 but non-trivial. Fan and Chen's review focused on a correlational analyses rather than an intervention effect summary.

These correlation meta-analyses have certainly painted a fuller picture of the nature of parent involvement in improving the academic performance of children. However, two recent intervention effect meta-analyses were reported that provide an ongoing summary and discussion of the intervention effects of parent involvement programs (Jeynes, 2003; 2005). In the 2003 review, Jeynes examined the effect of various types of parent involvement on academic outcomes for minority students. Parent involvement was broadly defined as ranging from parent expectations of academic success to parents attending or participating in school functions.

This important meta-analysis significantly advanced the understanding of the effects of minority parent involvement on student achievement from an intervention or programmatic perspective. Jeynes conducted an extensive literature search of some 25 databases that produced 26 potential studies of which 20 had useable data and were included in the metaanalysis. Four different measures of academic achievement were identified for the review including a composite of the three academic measures. Jeynes found substantial positive effects of parent involvement for African Americans (d=0.31), Latinos (d=0.48), and Asians d=0.22). While the author did provide discussion regarding validity issues for the cohort of studies, no specific analysis of the quality of design was conducted as part of the meta-analysis.

In 2005, Jeynes reported a second study of the effects of parent involvement programs in which a total of 41 studies were coded and analyzed to address four research questions focusing on the general question of the degree to which parent involvement can improve educational outcomes for urban children. This study provided a more detailed expansive analysis of parent involvement programs by addressing important methodological issues of design, controls, and transparency of effect size calculations. Further, he provided an analysis of differences associated with the more direct parent involvement programs compared to the voluntary or pre-existing programs. While this review did provide for an assessment of the impact of what he termed “sophisticated controls” a more detailed analysis of the experimental design characteristics would seem warranted.

The Need for a Systematic Review

The weight of evidence for parent involvement as a mechanism for improving academic performance in children has been generally accepted but difficult to assess. Narrative reviews have included bibliographic searches and annotated bibliographies. Most concluded that there was a positive result for children whose parents were involved in their academic learning. However, one may question the validity of these conclusions given that the reviewers must subjectively evaluate numerous study characteristics and outcomes. (Bronfenbrenner,1974; Fantini, 1979; Henniger, 1979). Certainly some summaries such as Mattingly, et al. (2002) have produced conclusions contrary to others in the field (Jeynes, 2002, 2005). In addition many of the primary studies reported in the literature were more qualitative than quantitative. Even among the quantitative studies, the evidence is not clear and consistent as to whether parent involvement has a positive, negative, or null effect on elementary school age children's academic achievement. The reason for the lack of clarity is that researchers often have relied on retrospective surveys, correlational studies of existing conditions, and anecdotal observations of both the quality of parent involvement and student achievement.

We cannot turn to the two meta-analyses on the topic to reconcile the mixed conclusions reported in primary studies, because of their methodological limitations. At best, these meta-analyses suggest that parent involvement has a positive effect on achievement.

The purpose of this review is to apply a set of a priori systematic procedures leading to a meta-analysis of studies with adequate controls to produce the least biased estimate of the effect of parent involvement on student achievement. To control for variation in threats to internal validity of a particular study design, we limit our review to studies that include at least two groups and use random assignment to form a fair comparison between groups. In other words, we limited our review to studies that use a randomized controlled trial (RCT) design.

Information Retrieval

Following the policies and guidelines in the Campbell Collaboration's Information Retrieval Methods Group (C2 IRMG) Policy Brief (2004), we considered the following information retrieval tasks and related activities:

Electronic Searches of Databases,

For the electronic searches, we documented our search strategies as recommended in the C2 IRMG Policy Brief by presenting tables that contain the following information:

Database Searched

The purpose of this documentation is to enable other researchers to replicate our searches. Prior to searching, we consulted with Anne Wade, the Education Coordinating Group's Information Retrieval Liaison and Alan Gomersall and Lesley Grayson of the UK Center for Evidence Based Policy and Practice. They provided feedback on our target list of databases to search and on our search strategies used with these databases. A number of the databases on this list required special access, especially those outside the US. Anne Wade and Lesley Grayson searched those databases on our behalf.

Electronic Searches of Databases

We began with a search of 27 electronic databases. The purpose was to be as comprehensive as possible and thereby reduce bias in the search for evidence. Our information retrieval experts advised that we include databases most likely to catalog studies on parent involvement. We then modified the list by removing some databases and adding others. For each database, we used the following search keywords to locate potentially relevant studies for the review:

Intervention Keywords: 1 parent* Involvement, or parent* participation, or parents-as-teachers, or parent* effectiveness, or parent* effectiveness training, or parent-child relationship, or parent* education;

We modified these terms to conform to each database's thesaurus, when a thesaurus was available, and to ensure that the universe of appropriate synonyms had been used during the search. Search term syntax was adapted to the requirements of specific databases. The process of locating relevant studies for review was iterative. Thus, for some database searches the final search strategy contained far fewer terms than those presented above.

Table 1 presents the results of our search of 27 databases supplied by entities in the United States, Canada, China, and the United Kingdom. Our intent was to include databases supplied by entities in other countries such as France and Germany. However, access to these databases (such as FRANCIS and SOLIS) was restricted to fee-paying subscribers. Still, as Table 1 shows, our search was comprehensive enough to retrieve over 800 citations with PsycINFO and ERIC yielding the highest number of citations at 205 and 191, respectively. On average, our search strategy retrieved 30 citations per database.

We searched databases such as ERIC, Chinese ERIC and PsycINFO, and discovered that they catalog a significant amount of grey literature which consists of unpublished reports such as dissertations, organizational reports, and government documents. In addition, a number of databases that we searched, such as Social Science Research Network, Index to Theses, and Digital Dissertations only catalogued grey literature. Our ability to expand our searches to databases that catalog grey literature, especially those in Europe, was limited because most of these databases required a subscription.

Hand Searches of Journals

Parental involvement is an interdisciplinary topic and studies on it are published in a wide variety of journals. Originally, we planned to conduct hand searches of specific journals based on the number of randomized controlled trials published in a particular journal as identified through our searches. However, the majority of the randomized controlled trials were identified in the grey literature (dissertations specifically) for which hand searching is not possible. The randomized control trials we identified in academic journals were widely distributed across journals such that a hand search was neither an efficient nor a cost-effective use of human or economic resources.

Contacting the Invisible College

Previous research in the social sciences (Glass, 1981) and more recent research in the health sciences (Hopewell, Clarke, Lusher, Lefebvre, & Westby, 2002) have shown that including grey literature in a meta-analysis can reduce bias. How much bias is reduced will vary according to the amount and quality of the grey literature excluded from the review, and the review topic (Egger, Juni, Bartlett, Holenstein, & Sterne, 2003). For instance, unpublished studies reside in the file cabinets of researchers because the results of the study were not statistically significant. These, are particularly difficult to obtain without personal communication. To identify such studies, an email was sent to over 1,200 scholars, researchers, policymakers, and practitioners worldwide informing them of our review and requesting a referral to a potentially relevant study, a person who might know of such studies, or an organization that sponsors, conducts, or has an interest in parent involvement studies.

Seventy members of the invisible college responded to the email campaign. Of these, 17% (12/70) referred us to a study; 10% (7/70) referred us to an organization; and 70% (49/70) referred us to a researcher. Ten percent expressed interest in the review and requested a copy of the published version.

When respondents emailed us the full-text for a study, we included the study in the initial screening pool and reviewed the reference list for additional leads. When respondents emailed us a citation, we searched the appropriate database to retrieve the abstract and if it was relevant retrieved the full text for further screening. To follow up on referrals to organizations and researchers, we sent the same email and followed the same process as when studies were referred to us.

None of the referrals led to our identifying a randomized controlled trial. However, we did collect a list of researchers, policymakers, and practitioners that expressed an interest and requested a copy of the final review.

Contributing to C2-SPECTR

All citations of studies identified during the information retrieval phase of this review and identified as a RCT or systematic review (that could have references to RCTs) were denoted as such with an asterisk in the tables of included and excluded studies. Studies not relevant to this review might be relevant to other C2 reviews and therefore are relevant for inclusion in the Campbell Collaboration's web-based Sociological, Psychological, Educational, and Criminological Trials Register (C2-SPECTR). As recommended in the IRMG Policy Brief, these tables will be forwarded to the person in the Campbell Collaboration currently responsible for cataloging studies in C2-SPECTR.

Design Characteristics

We evaluated the following characteristics of included randomized controlled trials (RCTs) in the review based on the following criteria if reported:

Method of random assignment

Two reviewers evaluated each study for attrition rates. Attrition rates were calculated based on a reduction of the number of participants from the pre-intervention to the post intervention measures. Studies were coded for the percent of participant attrition for both intervention and control groups.

Participant Characteristics

Each study reported data on a minimum of five participants in both experimental and control (or comparison) groups. The minimum number of group participants was selected in order to maintain the interpretability of the mean and standard deviation that were used in effect size and standard error calculations. Specific child participant characteristics for both the experimental and control (or comparison) groups were coded to include the following:

Age

Intervention Characteristics

All intervention programs were implemented for a minimum of 20 school days (or four weeks) as this was considered to be a potentially minimally meaningful length of time that a school based program might be implemented and reflect substantive results. No studies were excluded due to length of the intervention program. All intervention programs involved planned and consistent direct contact between the parent(s) and child in an educational activity outside the school-day setting. Studies in which parents were engaged in a classroom activity as an aide or support personnel were excluded from the review. In addition, intervention programs that measured parent involvement for participation in school-related activities such as parent teacher organization meeting, school program attendance, or home visits were also excluded from the review.

We chose a focused inclusion and exclusion criteria for the intervention program for several conceptual reasons. First, it is the case that many parents work during the school day. A comparison of parent involvement activities occurring during the school day could potentially confound outcome interpretation of studies in which parents were not able to participate during schools hours. Secondly, we were interested in the direct impact of parent involvement on achievement and we viewed supportive activities as an indirect intervention program. We recognize that these ‘support’ activities are potentially important and viable strategies for improving children's academic performance and warrant further consideration.

The following approaches to parent involvement were identified in this review.

Collaborative reading: Parents and children reading together as a structured activity.

Outcome Characteristics

Outcome measures for each intervention program contained a quantified assessment of educational performance using norm reference tests, criterion referenced tests, or rating scales of academic performance with appropriate evaluation of each measure's psychometric properties.

Timing of Outcome Measurement

Outcomes were measured at various points in time after the intervention was completed. Only those outcomes reported immediately post-intervention were included in the calculation and synthesis of the intervention's effect.

Homogeneity Analysis

A homogeneity analysis examined whether the variation in a set of effect sizes may be attributed to sampling error alone or to other factors. Data synthesis was conducted using both a fixed and random effects models and report the results of both were reported but rely on the random effects model data was relied on to interpret findings. Under the random effects model, the estimates vary from study to study because of differences among the study population parameters (between-studies variation) and the sampling of different subjects within the study populations (within-study variation). Results from the random effects model allow for inferences to the population of studies from which the set was sampled in which case results of the data synthesis can be extrapolated beyond the studies included in the meta-analysis (Kline, 2004).

Sensitivity Analysis

We tested the robustness of our meta-analytic results to any one study in the meta-analytic portfolio of studies through a sensitivity analysis. The one study removed analysis assessed how the effect size changes with one study removed relative to the average effect size with all remaining studies included.

Publication Bias

To assess publication bias, we used the trim and fill procedure and visually inspected the resulting Funnel Plot (Rothstein, Sutton, and Borenstein, 2005).

Incomplete Reporting of Study Data

RCTs that met our inclusion criteria reported enough information to compute effect sizes. If information reported in an RCT was insufficient to compute an effect size, we attempted to contact the author(s). For example, we contacted the author of one study but the author was unresponsive. In a second instance, we were unable to locate the author or anyone who could assist us. In both cases, we excluded the studies from the analysis, rather than replacing the missing effect size value with zero which Pigott (2001) has shown to be an inappropriate option. This option tends to underestimate the effect size variance and can lead to incorrect inferences.

Post Hoc Subgroup and Moderator Analyses

As part of the meta-analysis, we conducted a sub-group and moderator analysis to examine differences in the average effect sizes by the following study-level characteristics: 1) Method of Parent Involvement and 2) Publication Type. Meta-Regression was used to assess the relationship between length of intervention and effect size.

Descriptive Statistics for Included Studies

Nineteen RCTs met the inclusion criteria for the systematic review. One study, Phillips (1990), did not advance to the analysis stage because of a mismatch between the unit of assignment (schools) and the unit of analysis (students). Because outcomes were analyzed at the student level but the students were nested within the randomly assigned classrooms, their achievement measures were not independent. The remaining eighteen studies were published between 1964 and 2000 with the majority published as dissertations (67%). Most of the lead authors were female (83%).

The length of the parent involvement intervention programs ranged from 4 to 104 weeks with a mean of 23.2 weeks and a median of 10.5 weeks. The discrepancy between the mean and median was due to 1 study reporting 104 weeks of intervention. The most frequently reported method of parent involvement was collaborative reading (n = 8) followed by education and training (n = 4).

Of the studies that reported the socioeconomic status of the parents, 73% were from mixed SES, 9% were from middle SES and 18% were from lower SES. Of the studies that reported the ethnicity of the child, 25% were African American, 25% were Caucasian and 50% were of mixed ethnicities. Most of the intervention programs involved a mix of primary grades (47%) while 11% involved kindergarten, 11% involved first grade and 21% involved second grade; the remaining 10% were not reported.

All but one study was conducted in public schools (one study's school type was unreported) with twelve studies reporting a school location of suburban, urban, or rural. The majority (85%) were conducted in the United States while 10% were conducted in the UK and 5% were conducted in Canada. Most studies exhibited no differential attrition (70%).

Meta-Analysis of Included Studies

Specific Effects of Parent Involvement

Post Hoc Subgroup and Moderator Analyses

The results reported to this point enabled us to assess the main effect of parent involvement on achievement in general as well as specifically for reading, math and science. The effect of parent involvement may, however, vary by study characteristics such as the method of parent involvement. In other words, they may be interaction effects (at the study level) that might be uncovered through a planned, post-hoc subgroup analysis. Only three potential moderator variables provided an opportunity to unpack for whom parent involvement make the most difference: Approach of Parent Involvement, Length of Parent Involvement, and Publication Type.

Length of Intervention Program

A meta-regression analysis is presented in Figure 10 and shows there is no evidence of a relationship between the length of parent involvement and effect size (

b ^ = 0.01

, 95% CI = 0.00 to 0.01). This means that parent involvement intervention programs that are longer in duration do not produce larger effects that parent involvement intervention programs that are shorter in duration. However, the pattern of the available data does not permit one to make an inference about the linear relationship between parent involvement and the length of the intervention program.

Effects of Parent Involvement

Overall. It is clear that parent involvement has a positive and significant effect on children's overall academic performance. It is also clear that the effect is large enough to have practical implications for parents, practitioners, and policymakers (d= 0.45). The overall effect suggested that when parents participated in academic enrichment activities with their children outside of school, the benefits were manifest in improved academic performance in school. This result was striking when one considers that the median length of parent involvement was only 11 weeks. These finding are generally consistent with the results reported by Jeynes (2005) which reported d=.31 for those studies he classified as direct parent involvement programs. Further, both Jeynes findings and those of this study argue against the conclusions of Mattingly, et al. (2000). While Mattingly et al. concluded that there is little effect of parent involvement on academic performance, this review has uncovered compelling support for the use of a parent involvement program as a viable supplementary intervention to improve children's academic performance in school, and for the parent involvement component of the No Child Left Behind mandate

Reading Achievement. Parent involvement produced a positive and significant effect on children's reading performance (d =0.41). This result suggested that when parents participated in reading related activities with their children outside of school, a significant improvement in reading performance was manifest in school. Given the concern for improved literacy among school age children, the data from this review supports the need for parents, practitioners, and policymakers to consider parent involvement in reading related activities with their children as a viable strategy to support school-based reading instruction.

Math Achievement. Parent involvement in math achievement produced the largest specific effect on children's academic performance (d = 0.54). However, this result was biased by the Heller (1993) study which produced an effect size that was three times as large as the effect size for any other study. When this study was removed from the analysis, the impact of parent involvement on math achievement was reduced by approximately 50% to a nonsignificant effect. The results do not support the use of parent involvement as a recommended strategy to improve children's math achievement in school. However, the quality of the Heller study and the size of the impact of parent involvement on children's math performance were so large, for a single study, that it should not be overlooked given the concern with improving children's math achievement in the United States.

Parent Involvement Intervention Program. A goal of this review was to determine the parent involvement intervention program that produced the largest effect. Four approaches were investigated: Collaborative Reading, Education & Training, Games, and Rewards & Incentives. The largest impact of parent involvement on children's academic performance were produced by Rewards & Incentives (d = 1.18) and Education & Training (d= 0.61). However, the effect by parent involvement intervention program was confounded with other study characteristics such as children's grade level—the Rewards & Incentives intervention program was implemented with 4^th and 5^th graders, while the Education & Training intervention program was implemented with 1^st and 2^nd graders measured by reading outcomes.

Implications for Policy and Practice

This review concluded that there was substantial support for the positive impact of parent involvement with children outside of school to improve children's academic performance in school.

Implications for Research

There is a need to improve the quality of RCTs on parent involvement through adequate funding so that researchers can hire and train a sufficient number of research staff, recruit participants, and conduct fidelity of intervention and RCT implementation checks.

Implications for Reviewers

Publication Bias. Research has shown that studies published journals have a tendency to report larger effect sizes that studies published in grey literature. Our analysis of the 18 included studies revealed that their effect sizes differed by publication source. The average effect size of studies published in journals was almost twice as large as effect sizes from studies published in grey literature. This means that if we had retrieved studies from journals only, we would have over estimated the effect of parent involvement on children's academic performance by approximately 100%. Conversely, if we had retrieved studies from the grey literature only, we would have under estimated the effect of parent involvement on children's academic performance by approximately 50%. Therefore, these results support the need for a comprehensive search strategy to retrieve studies from both journals and the grey literature.

Review Limitations

Information Retrieval. We limited our search to English literature databases because of restrictions to electronic access, subscription fees, and language barriers.

Included Studies. The majority of studies (n = 12) came from dissertations. This limitation reflects less on the meta-analysis and more on the investments in the quality of RCTs conducted in the field of parent involvement.

Definition of Parent Involvement. Parent involvement is recognized as having many dimensions. We defined parent involvement narrowly (i.e., parents were required to actively engage with the child).

Definition of Target Population. This review focused on RCTs that assessed the effect of parent involvement on elementary school aged children. The nature of parent involvement changes with age and may require a qualitatively different type of parent involvement with the child. The outcomes are recognized as core learning skills fundamental to educational development. Finally, this review does not speak to the effect of parent involvement on children with special needs, such as dyslexia, etc.

Definition of Outcomes. This review focuses on educational achievement of the child only. It does not examine other important psycho-social outcomes for the child such as learning attitude, self esteem, and locus of control.

Review Maintenance

In accord with the IRMG Policy Brief, the authors will examine the review every 3 years for update. Reasons for updating or not updating the review will be documented and forwarded to the Education Coordinating Group liaison to the IRMG (as recommended in the IRMG Policy Brief),

Responsible Reviewer(s) for Update of Review

The first author will be responsible for updating the review

Acknowledgments

The authors would like to acknowledge Anne Wade, Lesley Grayson, and Alan Gomersall for their expert advice on information retrieval; Kelly Borradaile for assistance in organizing the manuscript; the Education Coordinating Group Co-Chairs for conference calls that refined our thinking about how to conduct a systematic review; students in our Systematic Review course at the University of Central Florida that did the same; Dorothy de Moya for editorial, financial and moral support; and Robert F. Boruch for his thoughtful editorial comments.

This protocol was begun while Chad Nye was a University of Pennsylvania/Campbell Collaboration/Robert Wood Johnson Post Doctoral Fellow at the University of Pennsylvania 2001-02.

Sources of Support

The Campbell Collaboration Secretariat provided a grant to complete editing of the final draft of the review.

Supporting Non-authors

None.

Statement Concerning Conflict of Interest

None known.