Indicated Truancy Interventions for Chronic Truant Students

Abstract

Objective: Examine the effects of indicated interventions on attendance with chronic truant students. Method: Systematic review and meta-analytic methods, following Campbell Collaboration guidelines, were utilized. A comprehensive search identified 5 randomized and 11 quasi-experimental studies. Results: The mean effect on attendance outcomes was moderate, positive, and significant, g = .46, 95% confidence interval [.30, .62], p < .05, translating into an improvement in attendance by an average of 4.69 days; however, significant heterogeneity was observed. Moderator analyses found no significant differences in mean effects between studies on variables tested. Discussion: Chronic truant students benefit from interventions targeting attendance behaviors; however, no program stood out as being more effective than others. Mean rates of absenteeism at posttest in most studies remained above acceptable levels, indicating a need for more effective interventions. The paucity of research, gaps, and deficiencies affirm the need for strengthening the evidence base. Recommendations for practice, policy, and research are discussed.

Keywords

truancy systematic review meta-analysis attendance absenteeism

Introduction

Truancy is a significant problem and major concern in the United States, United Kingdom, Canada, and other countries around the world. Truancy has been linked to serious immediate and far-reaching consequences for youth, families, and schools and communities, which has led to significant efforts being put forth by researchers, practitioners, schools, and policy makers to try to understand and address the problem. The United States and United Kingdom have implemented policies and invested significant resources to reduce absenteeism over the past several decades. Despite these significant efforts and millions of dollars spent, there is little evidence that any positive impact has been made on school attendance (Attwood & Croll, 2006; Davies & Lee, 2006). In the United States, the patterns of absenteeism remained relatively stable between 1994 and 2005, while the number of truancy cases petitioned and handled in juvenile courts in the United States increased 69% between 1995 and 2004 (National Center for Education Statistics, 2006; Stahl, 2008).

Truancy has been linked to significant negative implications for youth as well as for families, schools, and communities. The negative outcomes associated with truancy include additional delinquency, poor school performance, school expulsion and dropout, substance use, and other risky and problematic behaviors (Maynard, Salas-Wright, Vaughn & Peters, 2012; National Center for School Engagement (NCSE), 2007; Petrides, Chamorro-Premuzic, Frederickson, & Furnham, 2005; Reid, 1999). The long-term economic implications for students are also significant. Students who are chronically absent are more likely to perform poorly in school and more likely to drop out, which negatively impacts earning potential over their lifetimes (Attwood & Croll, 2006; Garry, 1996). The implications for schools with large absenteeism problems include loss of funds and failure to meet performance requirements (Goldstein, Little, & Akin-Little, 2003). Significant costs to communities associated with truancy and absenteeism include higher rates of criminal activity, citizens not productively contributing to the community, and higher government spending for social services (Baker, Simon, & Nugent, 2001).

Significant attention has been paid to the study of the causes and correlates of truancy, resulting in a large body of research linking numerous risk factors to truancy. Risk factors at the individual, family, school, and community levels have been found to demonstrate some relationship to truancy (Bowen, Bowen, & Ware, 2002; Corville-Smith, Ryan, Adams, & Dalicandro, 1998; Malcolm, Wilson, Davidson, & Kirk, 2003). This accumulated body of knowledge has led to truancy being increasingly recognized as a serious and complex problem influenced by multiple and interacting factors (Kearney, 2008; Kim & Streeter, 2006; Lauchlan, 2003). To reduce truancy and increase attendance, researchers, practitioners, and schools have developed various strategies targeting a number of risk factors associated with truancy, resulting in diverse interventions and policies being implemented in various settings.

Interventions to Increase Student Attendance or Reduce Truancy

Interventions targeting school attendance are diverse and fall into several different categories, target a variety of different risk factors and levels, are implemented in different settings, and are delivered through a variety of modalities (see Kearney, 2008; Reid, 2002). The number and types of interventions designed to increase student attendance or reduce truancy has been growing substantially. The Office of Juvenile Justice and Delinquency Prevention (OJJDP) established the Model Programs Guide, a database of programs that have met OJJDP’s methodological standards and have demonstrated effectiveness in impacting a number of different problems of concern by OJJDP. Sixteen truancy interventions are listed in the OJJDP Model Programs Guide. In addition, the NCSE has registered 171 truancy programs in their database, 69 of which have self-identified as having had an external evaluation and 30 of which have completed final evaluations. Despite the widespread attention to truancy and the increase in the number and variety of interventions available and in use to prevent and reduce truancy, truancy remains a significant problem.

Prior Reviews of Truancy Interventions

A number of prior reviews have synthesized knowledge on truancy and interventions to combat truancy and improve attendance. Most of these reviews have been narrative in nature and have not presented their findings systematically, often reviewing the same programs, emphasizing particular studies considered “effective” rather than computing effect sizes of the studies the reviews found, and failing to examine variability in effects by study, participant, or intervention characteristics.

Klima, Miller, and Nunlist (2009) conducted a meta-analysis of 22 experimental and quasi-experimental studies evaluating the effects of dropout and truancy interventions. Because the meta-analysis was part of a larger investigation of Washington State truancy laws, the focus of the review was on programs that could be implemented by at least one system involved in the larger investigation (schools, courts, or law enforcement) and thus excluded studies of programs carried out in elementary schools, social service, mental health, and other nonprofit organizations and programs for populations at risk due to minority or socioeconomic status (SES) and delinquency. In addition, the authors did not conduct moderator analyses to examine potential variation in effects due to study, population, or intervention characteristics. The authors reported small positive impacts on dropping out, achievement, and attendance/enrollment. For attendance and enrollment outcomes, the authors reported that alternative education programs, behavioral programs, and school-based mentoring programs were the modalities found to be most effective.

Sutphen, Ford, and Flaherty (2010) conducted a systematic review of the effects of truancy interventions. Their review included 16 studies of truancy intervention studies published in peer-reviewed journals between 1990 and 2007. The review included experimental, quasi-experimental, and single group pretest–posttest studies and a broad range of intervention modalities, including universal, selective, and indicated programs. The authors presented the findings for each study, but did not quantitatively synthesize them or calculate effect sizes for the included studies. The authors essentially used a vote-counting method and reported whether the primary study authors found significant differences between the treatment and intervention groups. The authors of the review mention a paucity of truancy intervention research and a lack of consistency in definitions of truancy used by researchers. They identified individual interventions that demonstrated beneficial effects, including interventions using contingency management, group guidance, and parental notification as well as some community-based and collaborative interventions.

The remaining reviews identified in the search were traditional narrative reviews of the literature. They examined the causes, correlates, and diagnostic features of truancy, highlighted various treatment modalities, and cited published intervention studies to provide evidence of the effectiveness of various treatments. In addition, the reviews were not systematic; they did not specify their search strategy or inclusion/exclusion criteria, and they included only published studies. Many of the narrative reviews focused on the same literature base, yielding considerable repetition.

All but one of the existing reviews utilized a narrative or vote-counting approach, presenting a description of programs or categorizing outcomes of programs as significantly positive, significantly negative, or neutral. Conclusions regarding effective interventions were then made based on the number of studies that were found to demonstrate significant positive results. The vote-counting method, however, disregards sample size, thus leading to erroneous conclusions (Glass, McGaw, & Smith, 1981). Also, the vote-counting method relies on statistical significance and does not take into account measures of the strength of the study findings, thus also leading to misleading conclusions (Glass et al., 1981). Meta-analysis, on the other hand, represents key findings in terms of effect size rather than of statistical significance. Moreover, meta-analysis provides a more transparent and valid analysis strategy than the alternative means of narrative reviews and vote-counting methods and provides a means to more systematically uncover gaps in the knowledge base (Lipsey & Wilson, 2001; Valentine, Pigott, & Rothstein, 2010).

From prior reviews, extant literature, and lists of truancy programs listed with the OJJDP, NCSE, and the National Dropout Prevention Center (NDPC), there appears to be a number of diverse programs that have been evaluated, providing a substantial body of research available for assessing the effectiveness of interventions to increase student attendance. Unfortunately, this knowledge is disparate and confusing, making it difficult for policy makers and practitioners to use the evidence of effectiveness to guide policy and practice. Although several narrative reviews and one meta-analysis of attendance and truancy interventions have attempted to summarize the extant research, there are a number of limitations to these reviews. This review improves upon prior work by systematically and statistically synthesizing truancy intervention research to provide a comprehensive picture of the effects of indicated interventions with chronic truant youth.

Purpose of the Present Study

The objective of this review is to examine the effects of indicated intervention programs on attendance outcomes of elementary and secondary school students who were identified as having chronic attendance problems. The specific questions guiding this review were: (1) Do indicated programs with a goal of increasing student attendance affect school attendance behaviors of elementary and secondary students who have an identified attendance problem? (2) Are there differences in school-based, community-based, and court-based programs with regard to services provided and effects on student attendance? (3) Are some modalities (i.e., individual, family, group, multimodal) of interventions more effective than others at increasing student attendance?

Method

Systematic review methodology was utilized for all aspects of the search, selection, and coding of studies. Meta-analysis was utilized to quantitatively synthesize effects of interventions and examine potential moderating variables. We followed the Campbell Collaboration procedures and guidelines for systematic review and meta-analytic methods (see www.campbellcollaboation.org). The protocol and data extraction form for this review are available in the review protocol from the Campbell Collaboration (see Maynard, McCrea, Pigott, & Kelly, 2009).

Eligibility Criteria

Types of studies

Studies utilizing randomized controlled trial (RCT) or quasi-experimental designs (QED) with a comparison group that received treatment as usual, no treatment or wait-listed, or an alternative treatment were included in this review. Single group pre–post test studies were also included, but analyzed and reported separately, in the full Campbell Review (Maynard, McRea, Pigott, & Kelly, 2012).

Types of participants

Students attending primary or secondary educational institutions and were identified by the original study authors as being truant or having an attendance problem. Studies in which participants were identified as school refusers for which their absenteeism was a result of anxiety or distress, were excluded from this review.

Types of settings

Any setting that serves primary or secondary school students was eligible. Studies conducted in residential facilities or psychiatric day programs were excluded from this review.

Types of interventions

Interventions with a stated primary goal of increasing student attendance (or decreasing absenteeism or truancy) specifically targeting truant/chronic absentee students who were identified prior to the study as having an attendance problem that met a threshold determined by the study researcher were eligible.

Types of outcome measures

School attendance was the primary outcome of interest in this review. Studies must have included at least one quantifiable measure of school attendance or absence and provided adequate data to calculate an effect size. Other outcomes were coded (i.e., school performance, anxiety) during the coding process; however, there were too few studies measuring the same secondary outcomes to conduct any meaningful analysis.

Geographical contexts

Due to significant differences in educational and legal systems around the world, this review included studies conducted in the United States, Canada, the United Kingdom, and Australia. Only English language articles were included in the review.

Time frame of field trials

This review includes studies that were published between 1990 and April 2009, even though the research itself might have been conducted prior to 1990. Increased attention to attendance problems and national initiatives to combat attendance problems and truancy began occurring in the 1990s in the United States, which resulted in a large number of evaluation studies assessing the effectiveness of attendance interventions. Therefore, this review focused on the past 20 years to provide a comprehensive and contemporary view of attendance interventions.

Search Strategy

A comprehensive search strategy was conducted in an attempt to identify and retrieve all relevant studies, both published and unpublished, which met the search criteria described above. Although this review is limited to indicated intervention programs serving students with an identified attendance problem, the search process was conducted to find universal and selective programs as well to identify studies that will be used in future reviews. Several sources were used to identify eligible studies, including:

Electronic databases

A total of 18 databases were searched (see Table 1). Two librarians specializing in social work, criminal justice, and education as well as experts through the Campbell Collaboration were consulted to determine the appropriate databases to search and keyword search terms to utilize.

Table 1.

Electronic Databases Searched

Name of Database
Academic Search Premier	Francis
CBCA education	MEDLINE
Canadian Research Index	PsycInfo
Cochrane controlled trial register	Questia
Criminal justice abstracts	Social Science Citation Index
Databases of abstracts of reviews of effectiveness	Social Service Abstracts
Dissertation abstracts	Social Work Abstracts
ERIC	Sociological Abstracts
Education complete	WorldCat

Keyword searches within each database included combinations of keywords with appropriate wildcards grouped into four main categories: (1) Outcome: Attendance OR Absence AND (2) Intervention: Evaluation OR Intervention OR Treatment OR Outcome OR Program AND (3) Targeted behavior: Truancy OR School refusal OR Absence OR Attendance OR School phobia OR School anxiety OR Dropout OR Expulsion OR Suspension AND (4) Targeted population: Students OR Schools

Internet and website searches

Websites of relevant government agencies, research centers, foundations, and professional associations were searched for published and unpublished studies. These sites included the U.S. Department of Education, the OJJDP, coloradofoundation.org, hfrp.org, truancyprevention.org, drgonline.com, Colorado.edu/cspv/blueprints/, schoolengagement.org, dropoutprevention.org, ies.ed.gov/ncee/wwc/, and Google Scholar.

Personal contacts

Personal contacts with research centers, organizations, and researchers who do work in the fields of truancy, school refusal, and school absenteeism were contacted. An e-mail query of researchers and experts in the field was attempted in an effort to uncover additional published or unpublished studies relevant to the review. In addition, efforts were made to contact all truancy and attendance programs listed on the NDPC and NCSE websites as well as programs listed in Reimer and Dimock’s (2005) booklet. Contact was attempted via e-mail inquiry to the contact person listed for the program. If no response was received from the e-mail inquiry or the e-mail came back as undeliverable, a letter was mailed to the contact person.

Reference lists

Reference lists of prior reviews and related meta-analyses were reviewed for relevant studies. In addition, the references of the retrieved primary studies were examined for studies potentially relevant for the review. Reference lists of prior reviews and retrieved primary studies yielded 11 studies that were retrieved and screened for eligibility.

Retrieval, Selection, and Coding of Studies

Titles and abstracts of the studies found through the search procedures were screened for relevance by the first author, and those that were obviously ineligible or irrelevant were screened out. For example, studies that were deemed inappropriate at the title/abstract review stage were those that did not involve the target population (e.g., they involved college students or adults), did not involve an intervention, or were theoretical in nature. If there was any question as to the appropriateness of the study at this stage, the full text document was obtained and screened. Documents that were not obviously ineligible or irrelevant based on the abstract review were retrieved in full text and screened for eligibility utilizing a screening instrument (see Maynard et al., 2009).

Studies of indicated programs that met the eligibility criteria were coded using a data-coding instrument developed by the first author. The coding instrument used for this review was comprised of five sections: (1) source descriptors and study context; (2) sample descriptors; (3) intervention descriptors; (4) research methods and quality descriptors; and (5) effect size data.

All study coding was done on a hard copy of the coding form and entered into Excel. Data needed for the meta-analysis were entered into Comprehensive Meta-Analysis (CMA, version 2.0; Borenstein, Hedges, Higgins, & Rothstein, 2005). All coding was completed by the first author. A random sample of 20% of the studies was coded by the third author or a trained research assistant. There was less than 10% discrepancy between coders in critical fields (data related to effect size and study design and quality). If there had been more than 10% discrepancy between coders, the remaining 80% of the studies would have been coded by a second coder and all discrepancies resolved.

Statistical Procedures

Statistical analysis was designed to produce descriptive information on the characteristics of the studies included, the effect size of each intervention on attendance outcomes, the grand mean effect size, the heterogeneity of effect sizes around the mean, and the relationship between effect sizes and methodological qualities as well as substantive characteristics of the samples and interventions.

The intervention outcome of interest for this review was attendance, which was reported as a continuous variable in all studies. Attendance was measured and reported in terms of mean number of days attended or absent, mean number of classes absent, or mean percentage of days attended or absent. Effect sizes were calculated in CMA version 2.0 (Borenstein et al., 2005). The standardized mean difference effect size statistic was utilized for the RCT/QED studies, employing Hedges’ g to correct for small sample size bias (Hedges, 1981). In cases where the authors did not report the means and standard deviations needed to calculate an effect size, but did report the results of a t test or one-way analysis of variance (ANOVA), the effect size was calculated in CMA by inputting the means, sample sizes, and t value, or in the case of an F ratio, the sample sizes and square root of the F ratio. In cases where reported data did not allow for the calculation of effect sizes, and it was not possible to estimate the effect sizes with values from t tests or ANOVAs, the study was excluded.

To maintain statistical independence of data, only one effect size was computed for each subject sample. In cases of studies with more than one treatment group, the group that was deemed most relevant was included in the meta-analysis. In cases of studies with more than one comparison group, the comparison group that received the least amount of intervention was utilized. For studies that reported attendance/absence in more than one way (i.e., reported average attendance by full day absent and total number of classes missed), the outcome that was most similar to the other studies included in the review was utilized.

A test of homogeneity (Q test) was conducted to compare the observed variance in the distribution of effect sizes to what would be expected from sampling error. The Q statistic is distributed as a χ² with k–1 degrees of freedom (k = the number of effect sizes; Hedges & Olkin, 1985). A significant Q rejects the null hypothesis, indicating that the variability in effect sizes between studies is greater than what would be expected by sampling error alone.

Moderator analysis was warranted due to the heterogeneity of effect sizes between studies being larger than expected from sampling error alone (details in the Results section). Random effects models were used for all analyses. The analog to the ANOVA was employed to test the association between categorical independent variables and variability in the effect sizes. Bivariate meta-regression was employed to test the association between continuous variables and variability in the effect sizes. The independent variables tested for moderating effects were: study design, publication type, attrition, grade level, type of intervention, treatment duration, modality of treatment, student grade, race, and chronicity of absenteeism at baseline.

Publication bias was also assessed in this review. Publication bias can occur as a result of decisions on the part of authors as well as editors to publish studies that demonstrate a significant effect and to not publish studies when findings may be insignificant, or run counter to the hypothesis or conventional wisdom (Cooper, 1998). Including only published studies in a meta-analysis could likely introduce an upward bias into the effect sizes (Lipsey & Wilson, 2001). Therefore, it is recommended that meta-analysis include both published and unpublished studies to minimize this bias (Cooper, 1998; Lipsey & Wilson, 2001). This review made every attempt to include both published and unpublished reports to minimize the occurrence of publication bias. In addition, publication bias was assessed by constructing a scatter plot of the effect size by sample size, called a funnel plot.

Results

The search yielded over 8,700 “hits.” After review of titles and abstracts, a total of 694 studies were retrieved for screening, with 391 of them meeting basic criteria as an attendance intervention. Those 391 studies were then categorized into type of intervention: 88 studies were categorized as universal interventions, 239 studies were categorized as selective interventions, and 64 studies were categorized as indicated interventions. Of the 64 indicated intervention studies, 5 RCT, 11 QED, and 12 SGPP studies met inclusion criteria for the full Campbell review. Studies were excluded at this stage due to authors not providing sufficient information to calculate effect sizes or because the studies were evaluating interventions targeting school refusal or school phobia. See Figure 1 for the flowchart detailing the search and selection process. A list of included studies and excluded studies with reasons for exclusion can be found in the full review (see Maynard, McCrea, et al., 2012).

Figure 1.

Study search and selection process flowchart.

Descriptive Analysis

The search strategy identified 28 eligible studies reported in 26 individual reports. Five of the studies were RCTs, 11 were QED studies, and 12 were single group pre-post test studies (SGPPs). For the purposes of this report, the results of the analysis including only the RCT and QED studies will be reported.

Table 2 summarizes the characteristics of the included studies. Despite efforts to search for studies conducted in a broad geographical area—including the United Kingdom, Australia, and Canada—all of the included studies were conducted in the United States. The comparison condition for the majority (88%) of the studies was treatment as usual, no treatment, or waitlist. Two of the studies compared the treatment condition to a comparison group that received an alternative intervention.

Table 2.

Characteristics of Included Studies

Characteristic	N (%)	Characteristic	N (%)
Publication Year		Country
1990–1999	11 (69)	United States	16 (100)
2000–2009	5 (31)	United Kingdom	0 (0)
Publication type		Author profession
Journal	4 (25)	Social Work	3 (19)
Dissertation or thesis	10 (63)	Psychology	3 (19)
Other report	2 (12)	Education	7 (44)
Sample size		Nursing	1 (6)
1–29	2 (13)	Criminal Justice	0 (0)
30–59	5 (31)	Unknown	2 (13)
60–99	4 (25)	Comparison group condition
100–199	2 (13)	Nothing or treatment as usual	14 (88)
200+	3 (19)	Alternative intervention	2 (13)
Author involvement in intervention
Yes	5 (31)
No	7 (44)
Unsure	4 (25)

Of the 16 studies included in this meta-analysis, 75% were found in the gray literature. The unpublished studies included 10 dissertations, theses, or master’s research papers and two reports by a school district. Sample sizes of the included studies were fairly small. The mean sample size of the included studies was 54 (range 5–193; SD 65.4). Although attrition was not a problem in the majority of the studies (88%), several studies reported challenges with obtaining or retaining larger samples even though they had originally planned for more participants. Some of the challenges cited by authors included: (1) problems locating and connecting with students and parents to enroll them in the study due to the high mobility of the families; (2) difficulties obtaining current residency and contact information from the school system; (3) disengagement and lack of trust in the school system that contributed to families’ reluctance to be contacted or to give consent for participation in the study; (4) challenges in obtaining complete attendance records for students leaving the school system; and (5) participant dropout from treatment or control conditions.

Attendance was measured as a continuous variable in all studies, and attendance data were obtained from an official school record or verified against an official record. Authors varied in the ways they operationalized attendance or absence and in how they reported attendance outcomes. In terms of authors’ operationalization of absences, some authors measured only unexcused absences, some utilized both unexcused and excused absences, and some factored in tardies or partial days absent while others utilized only full days absent. Some authors were not transparent about what they were including in their reported absence rates. In terms of the formats authors used to report absences, some authors reported attendance rather than absences and did so in terms of mean number or mean percentage of days. All authors used the same method for measuring and reporting attendance/absence for their treatment and comparison groups. Two authors reported attendance data in two different ways, of which only one form of data from each study was utilized to calculate the effect size.

Participants

A total of 1,725 students participated in the studies; 902 students received the treatment condition and 823 received the comparison condition. Studies included in the meta-analysis targeted a range of grade levels. Elementary students were the target of intervention in 20% of the studies, middle school students in 17% of the studies, high school students in 17% of the studies, and a mixture of grade levels in 46% of the studies. The majority of studies did not adequately report the race of the participants or the SES of the participants. For those studies that did report race (n = 12), Caucasian participants were the predominant race in 33% of the studies. Although students with attendance problems were targeted in this review, the degree of severity of absence rates varied across studies. Rates of absenteeism for students at pretest ranged from a mean of 9% to over 40%. Table 3 summarizes the characteristics of the participants of the included studies.

Table 3.

Participant Characteristics

Characteristic	N (%)	Characteristic	N (%)
Mean age^a	13.17	Predominant race
Grade level		Caucasian	7 (33)
Elementary	2 (10)	African American	3 (14)
Middle school	5 (24)	Hispanic	2 (10)
High school	5 (24)	Not given	9 (43)
Mixed grades	9 (43)	Socioeconomic status
Rates of absenteeism^b		Low	3 (14)
<10%	1 (5)	Working class	1 (5)
11–20%	1 (5)	Not given	17 (81)
21–30%	3 (14)
31–40%	4 (19)
41%+	7 (33)
Not given	5 (24)

Note. Reported for treatment groups.

^aAge reported in 13 studies.

^bPretest mean rates of absenteeism.

Interventions

The interventions in this review represent a broad range of program types and intervention modalities provided by a range of providers from different disciplines (see Table 4). The majority of interventions were ongoing, with one intervention being a single event that occurred in 1 day. The duration of the interventions ranged from 1 to 72 weeks, with a mean of 18.8 weeks (n = 13). Each intervention was categorized into one of the following types of programs: (1) school-based interventions; (2) court-based interventions; and (3) community-based interventions (see Table 4) and individual intervention components were identified (see Table 5). Because several studies used more than one component, the categories are not mutually exclusive.

Table 4.

Intervention Characteristics

Characteristics	N (%)	Characteristics	N (%)
Program types		Providers^a
School based	12 (75)	Social worker	5 (31)
Court based	3 (19)	Psychologist	3 (19)
Community based	1 (6)	Counselor/therapist (unspecified)	2 (13)
Settings		Counselor, school	1 (6)
School	11 (69)	Teacher/school staff	10 (63)
Court	1 (6)	Court staff	1 (6)
Clinic/agency	0 (0)	Peers	1 (6)
Multi/varied	4 (25)	Multiple providers	7 (44)
Unable to determine	0 (0%)	Treatment duration (N = 13)
Collaborative		One event	1 (6)
Yes	3 (19)	1–4 weeks	0 (0)
No	13 (81)	5–9 weeks	3 (19)
Unable to determine	0 (0)	10–18 weeks	6 (38)
		19+ weeks	3 (19)
		Unknown	3 (19)

^aCategories not mutually exclusive.

Table 5.

Components of Interventions

Student-Targeted Interventions	N (%)	Parent/Family-Targeted Interventions	N (%)
Counseling, social work, other therapeutic intervention	5 (31)	Family therapy	2 (13)
CBT—Individual	0 (0)	Educational group meetings	1 (6)
CBT—Group	1 (6)	Interdisciplinary team meetings/conferences	1 (6)
Group Therapy (non-CBT)	4 (25)	Criminal prosecution	0 (0)
Individual therapy (non-CBT)	1 (6)	Referrals for services	1 (6)
Behavioral interventions (contracting, incentives/rewards)	4 (25)	Parenting skills/training	1 (6)
Mentoring	3 (19)
Court proceedings	1 (6)
Pharmacotherapy	0 (0)
Individualized plans	1 (6)
Student health center services	1 (6)
Alternative education programs	3 (19)
Positive behavioral supports (PBS)	1 (6)
Case management	1 (6)
Peer support	2 (13)

Note. CBT = cognitive behavioral therapy; PBS = positive behavioral supports.

Categories are not mutually exclusive.

Table 6.

Mean Effect Size and Heterogeneity Statistics

	Hedges’ g	95% CI	k	Q	I ²	τ²
Attendance outcomes	.46	[.30, .62]	16	43.04*	65.13	.044

Note. *p < .05.

Mean Effects of Interventions

The overall mean effect size for attendance outcomes from the 16 independent samples reported in the 15 reports, assuming a random effects model, was .46, 95% confidence interval [.30, .62], p < .05, demonstrating an overall positive and statistically significant effect of interventions on attendance outcomes (see Table 6). The estimate of the random effects variance component was .044 and differed significantly from zero (Q = 43.04, p < .05). Table 7 provides a summary of the characteristics and mean effect sizes for each of the included studies. The mean effect size and confidence intervals for each study are also shown in the forest plot in Figure 2.

Figure 2.

Forest plot of mean effects of included studies.

Table 7.

Summary of Included Studies

Author (Year)	Program Name	Description/Components	QED/RCT	N	% Days Absent		Grade Level	Study Result	ES	95% CI Lower Upper
Author (Year)	Program Name	Description/Components	QED/RCT	N	Pre	Post	Grade Level	Study Result	ES	95% CI Lower Upper
Converse & Lignugaris/Kraft, (2009)	School-based mentoring program	Mentoring by school staff/faculty—once per week over 18 weeks	RCT	tx =16 c = 15	9	7	2	ns	0.56	[−0.14, 1.26]
DeSocio et al. (2007)	Truancy intervention pilot project	Student enrollment in a school-based health center for comprehensive health services and recruitment of teachers from within the students’ school to engage in mentored relationships	RCT	tx = 28 c = 31	32	63	62	+	0.51	[−0, 1.03]
Fantuzzo, Grimm, and Hazen (2005)	Project Start	Truancy court—courtrooms within designated school buildings (rather than traditional courtroom); caseworkers from service organizations located in the truants’ community present to promote family utilization of community services; referrals or direct services provided to families depending on their capacity of the caseworker	QED	tx = 189 c = 189	23	13	4	+	0.48*	[0.28, 0.69]
Flanagan (2006)	Going to Class Pays	Positive behavior support program—engaging in positive verbal interactions; utilizing attendance monitoring, positive parent interactions, and preferred reinforcements.	QED	tx = 32 c = 32	46	43	3	+	0.81*	[0.31, 1.32]
Glover (1990)	Group intervention and peer support	Social worker facilitated nontruant students in providing peer support in the context of group counseling for truant students. The social worker counseled the parents of the students in this group. Group met once per week for 30 min over 15 weeks	QED	Tx = 5 c = 5	ng	15	2	ns	0.57	[−1.12, 1.12]
Herrick (1992)	Incentive program for improved school attendance	Tangible incentives and verbal praise to students, who met with social worker weekly to develop contract, receive incentive/praise	QED	tx = 49 c = 15	25	13	1	ns	0.37	[−.021, 0.94]
Hess (1990)	Contingency contracting and parent training	Contingency contracts were developed with students. Parents attended 3 weekly group parent-training sessions	RCT	tx = 12 c = 15	49	23	2	+	1.05*	[0.22, 1.79]
Hess, Rosenberg, and Levy (1990)	Contingency contracting and group counseling	Contingency contracting—contracts developed with students and progress monitored daily. Group counseling—6 sessions over 10 weeks. Rational-emotive and theme-centered interactional approaches	RCT	tx = 13 c = 13	37	18	2	+	1.18*	[0.37, 1.99]
Hubin (2000)	Stop Truancy Project (SToP)	Information meeting held at courthouse. County attorney, social worker, and school rep discuss the legal, social, and educational ramifications of truancy, present on school and community resources	QED	tx = 15 c = 8	ng	ng	4	+	0.88*	[0.18, 1.58]
Johncox (1994)	School Success Project	Diversion conference with brief assessment, school attendance agreement signed by participants, referral for services (life management skills, in-home family counseling, psychological testing/evaluation). If further absences, re-staff and develop another plan, which may include court appearance	QED	tx = 45 c =17	27	20	4	ns	0.26	[−0.29, 0.817]
Johnson & Syropoulos (1996)-1	High School Intervention Centers Program	Students enrolled in three courses (language arts, mathematics, and group guidance); focused on individual needs of student in small group settings; intensive goal-directed guidance mode was used to promote self-awareness skills and effective problem solving	QED	tx = 193 c = 184	ng	28	3	+	0.82*	[0.61, 1.03]
Johnson & Syropoulos (1996)-2	High School Intervention Centers Program	Same as above	QED	tx = 165 c = 169	31	39	3	ns	0.06	[−0.16, 0.27]
Lawson (1990)	Peer tutoring	Peer tutors worked with tutees in 16 sessions, 30 min each (2×/wk for 8 weeks) covering preplanned topics/skills. The tutors (also truants) were trained by the PI, and biweekly meetings were scheduled with the tutees to address any problems	QED	tx = 60 c = 45	ng	11	1	ns	0.07	[−0.32, 0.45]
Seamans (1996)	Brief Family Systems Intervention	Six 1–1.5 hr family therapy sessions over 8 weeks	QED	tx = 34 c = 24	46	28	4		0.27	[−0.25, 0.79]
Sherriff (1990)	School-based special education program	16 weekly lessons in the Project Y classroom. Alternative classroom setting. Involves community work, personal and social development, work experience, and recreation	QED	tx= 14 c = 16	58	68	3	+	0.71	[−0.01, 1.43]
Tichenor (1991)	Making It in Middle School	Met with counselors as a group; positive reinforcement, problem solving	RCT	tx= 32 c = 35	15	14	2	ns	0.01	[−0.47, 0.478]

Note. ES = Effect size (Hedges’ g); QED = quasi-experimental designs; RCT = randomized controlled trial.

Grade level: 1 = elementary, 2 = middle school, 3 = high school, 4 = mixed grades.

Study results reported by author: + = author reported significant differences between groups; ns = author reported no significant differences.

*p < .05.

Analysis of Homogeneity

The test of homogeneity indicates whether the between-study variance is significantly different from zero. The result of the statistical test for homogeneity was statistically significant (Q = 43.04, p < .05), indicating that variability in effect sizes between studies was larger than expected from sampling error. Although the grand mean effect size provides evidence that the attendance interventions were, on average, moderately effective, the highly heterogeneous nature of the distribution suggests large differential effects across studies. Because the studies disagreed on the magnitude of effect, our next step was to further examine the reasons for this variability. The between-study differences in effects may be a result of factors associated with the study methodology or with participant or intervention characteristics. To explore the variability between studies and examine independent variables that may be contributing to the heterogeneity, moderator analyses were performed.

Analysis of Publication Bias

To mitigate potential publication bias, special efforts were made to search for and retrieve unpublished reports, resulting in 65% of the included studies being unpublished dissertations, theses, or reports. Due to the large number of unpublished studies as well as several small studies included in this meta-analysis, publication bias was likely mitigated. However, to formally assess the potential for publication bias, a funnel plot depicting the effect size (Hedges’ g) plotted against the standard errors was examined. The funnel plot, as shown in Figure 3, is reasonably symmetric, indicating that publication bias does not appear to be a source of bias in this review.

Figure 3.

Funnel plot of included studies.

Analysis of Moderator Effects

Study, participant, and intervention characteristics were tested in moderator analyses. Given the small number of studies in this review, we did not conduct any multivariate meta-regression models. The majority of moderator variables tested were categorical variables; therefore, moderator analysis for the categorical variables was performed using the analog to the ANOVA framework in which effect sizes were weighted by the inverse of the variance of the effect size estimate. Continuous moderators were examined with bivariate meta-regression models. The results of the moderator analyses of the categorical factors are presented in Table 8 and results of the meta-regression of the continuous factors are presented in Table 9. None of the moderators tested demonstrated a significant relationship with treatment effect. Although there were no significant differences in effects between studies on the moderator variables examined, some interesting findings are evident from the analyses.

Table 8.

Moderator Analysis Results for Categorical Factors

Factor	Q _b	K	Mean ES	95% CI
Study characteristics
Publication status	2.08
Published		5	.38	[.21, .56]
Unpublished		11	.55	[.38, .72]
Study design	.28
Random assignment		5	.57	[.17, .96]
Nonrandom (QED)		11	.43	[.28, .59]
Participant characteristics
Race	.43
Caucasian		4	.39	[−.10, .88]
African American		3	.45	[−.062, .96]
Hispanic		3	.55	[.23, .87]
Grade	1.59
Elementary		2	.16	[−.16, .48]
Middle school		5	.53	[.07, .99]
High school		5	.53	[.28, .78]
Mixed grades		4	.46	[.23, .68]
Intervention characteristics
Program type	.33
School based		12	.47	[.26, .68]
Court based		3	.49	[.20, .79]
Community based		1	.27	[−.25, .79]
Focal modality	.76
Group		5	.60	[.13, 1.07]
Family		4	.46	[.23, .68]
Mentoring		3	.28	[0, .56]
Alternative Ed.		3	.50	[.05, .94]
Contracting		1	.37	[−.21, .94]
Collaborative	.06
Yes		3	.49	[.30, .67]
No		13	.46	[.22, .70]
Multiple modalities	1.36
Yes		9	.54	[.31, .78]
No		7	.31	[.08, .53]

Table 9.

Bivariate Meta-Regression Results for Continuous Factors

Variable	k	b	SE	95% CI
Treatment duration	13	−.002	.007	[−.02, .01]
Baseline attendance rates	11	.013	.008	[0, .03]
Attrition rates	16	.002	.005	[−.01, .01]

Note. Some studies were not included in analyses due to not reporting sufficient data to code for the variable, thus n < 16 for some analyses.

Relationship of study characteristics to effect size

Variables tested for moderating effects related to study characteristics included publication status, study design, and rates of attrition. Of the study characteristics tested, none of the variables demonstrated a relationship to effect size. No significant differences in effects were associated with study design (Q _b = .28, p > .05). The inclusion of the weaker design (QED studies) did not appear to have upwardly biased the results, thus validating our decision to include QED studies in the analysis. Separating the QED studies from the studies that utilized randomization, or excluding them all together, would have served no purpose since they yielded essentially the same results (Glass et al., 1981). It should be noted that publication status and study design were highly correlated with each other. Studies that utilized random assignment were more likely to have been published. Published studies and studies utilizing randomized design were also correlated with studies in which interventions were tested with participants in middle school. Confounded moderators tend to introduce ambiguity in interpreting the results of univariate moderator analyses like those reported here. However, that none of the moderators exhibited a significant relationship with effect size magnitude gives us somewhat more confidence in our interpretations.

Relationship of participant characteristics to effect size

Variables related to participant characteristics for which moderator analyses had been planned a priori included baseline rates of absenteeism, grade, race/ethnicity, and SES. It is worth noting that only 10 studies reported adequate information on race or ethnic background and only 4 reported adequate information on the SES of the participants in the studies. Although race/ethnicity and SES are known to be correlated with absenteeism, authors did not regularly report data on the racial makeup or SES of study participants. Due to an insufficient number of studies reporting SES, moderator analysis could not be performed for this variable. For the 10 studies that did report race/ethnicity, no significant differences in mean effects were observed between studies comprised of samples with different racial/ethnic compositions.

Upon visual inspection, treatment effects for studies that included students with lower baseline absenteeism appeared to be smaller than studies that included students with higher rates of baseline absenteeism; however, no significant differences were observed, t(9) = 1.59, p > .05. No significant differences were observed in mean effects between grade levels (Q _b = 1.59, p > .05); however, only two studies were conducted with elementary students. In addition, 30% of the included studies were conducted with students across various grade levels. In studies that did include participants across grade levels, authors did not provide subgroup analysis by grade level to assess differential effects by grade. It should also be noted that grade level was highly correlated with both treatment modality and treatment duration. Studies involving high school participants were more likely to be alternative education programs and be longer in duration, middle school programs tended to use a group modality, and elementary programs tended to use behavioral contracting. Studies with participants from mixed-grade levels tended to be school- or court-based, employ family modalities, and be collaborative programs.

Relationship of intervention characteristics to effect size

Variables related to intervention characteristics examined for moderating effects included program type (school-, court-, or community-based); focal modality (group, family, mentoring, alternative education, and behavioral contracting); duration of treatment (number of weeks); collaborative interventions (yes/no); and multimodal interventions (yes/no). No significant differences in mean effects were found on any of the intervention characteristics tested.

Court-based, school-based, and community-based programs all demonstrated similar effects on attendance outcomes (Q _b = .33, p > .05). Interventions did not demonstrate statistically significant differences between the types of modality utilized (Q _b = .76, p > .05); however, mentoring, contracting, and alternative education interventions demonstrated effects that were not statistically different from zero within each of those groups. Similarly, no significant differences in effects were found between programs that utilized a single modality versus those that utilized two or more modalities (Q _b = 1.36, p > .05). Collaborative interventions did not demonstrate significantly larger effects than noncollaborative interventions (Q _b = .06, p > .05). The length of treatment also did not demonstrate a relationship to magnitude of effect, t(11) = −.26, p > .05 in the meta-regression. Shorter term interventions produced statistically similar effects to longer term interventions.

It should be noted that, due to the small number of studies in this review, most of the categories within the variables tested included a small number of studies, and in a few cases, only one study. For example, there was only one study of behavioral contracting. Due to the few number of studies and thus low statistical power, we may not have been able to detect moderator effects that may indeed be present. In addition, some intervention characteristics were highly correlated with each other as well as with participant and study characteristics. For example, collaborative programs were more likely to be school based, to be conducted with middle schoolers, and to have family as the focal modality of the intervention.

Clinical Significance

The overall effect size of attendance interventions examined was .46. We can translate this into terms that are more intuitively comprehensible by converting it back into number of days of school attendance that the treatment group gained as a result of receiving treatment. We selected all of the studies that measured and reported the mean and standard deviation of number of absences, which was the most common method used to report and measure attendance rates in the included studies. We then pooled the control group means and standard deviations for those studies into a grand mean and standard deviation using the procedures described by Lipsey and Wilson (2001). We then multiplied the effect size by the pooled standard deviation of the control group to calculate the number of days difference in attendance the .46 effect size represents. Following the above stated procedure, the .46 effect size for number of days absent translates into 4.69 days. That is, we can expect truant students who received an attendance intervention to improve attendance by 4.69 days.

Although improving attendance by 4.69 days, almost a full week, is important and most would agree is practically significant, the attendance rates reported at posttest in the majority of the included studies continued to remain above 10% (see Table 10). Although students who received an intervention did better on average than their control-group peers, students’ attendance did not improve to the point that they were achieving acceptable levels of attendance (if we assume attending school 90% of days or more is acceptable).

Table 10.

Posttest Mean Rates of Absenteeism

% of Days Absent	N	%
<10%	1	10
11–20%	7	40
21–30%	3	20
31–40%	1	10
41%+	3	20

Note. One RCT/QED study did not provide data in a way that enabled the % of posttest absences to be calculated. Several studies did not provide the exact number of school days for which they measured posttest absence/attendance, so assumptions were made in calculating the posttest absence rates. It was assumed that there are 180 days in a school year, 90 days in a school semester, 45 days in a marking period, and 5 days in a school week.

Conclusions

The literature on truancy is voluminous and disparate. Absenteeism research is spread across multiple disciplines, and much has focused on causes, correlates, and consequences rather than effects of interventions. This makes it challenging to know what, if anything, works to impact truancy. It also prevents practitioners and policy makers from using evidence to make decisions.

As indicated by the relatively few studies located in the search process, there is limited evidence on the effectiveness of truancy interventions aimed at increasing attendance for chronic truant students. The number and types of interventions currently in operation throughout the United States and other countries contrasts sharply with the number and types of interventions for which there are reasonably rigorous evaluations. It seems reasonable to conclude that the studies in this review do not adequately represent the outcomes of programs currently in existence and therefore cannot be generalized to the population of programs in operation.

Although there are relatively few studies in this meta-analysis compared to the number of programs currently in existence, these studies represent the best empirical evidence currently available for indicated truancy intervention outcomes. A comprehensive search for published and unpublished studies to include in this review yielded only 5 RCT studies and 11 QED studies that met inclusion criteria. Given that there is an abundance of literature documenting the causes, correlates, and negative impacts of truancy and absenteeism, and a general consensus that truancy is a serious issue, uncovering only 16 studies of outcomes of indicated interventions with truant students utilizing experimental or quasi-experimental methodologies is a concern. A number of interventions and programs have been recommended by experts, identified as effective or model programs, or listed in databases of national centers, which lend an air of credibility to these interventions. Despite this, the relatively small number of studies that met inclusion criteria indicates that there is scant evidence on the effectiveness of current programs in existence.

Overall, interventions included in this review were found to demonstrate a significant though moderate, positive effect on attendance outcomes. While the mean effects of the interventions were moderate and significant, it is important to note that the heterogeneity of effect sizes was also significant, indicating that different studies point to somewhat different conclusions. Additionally, none of the moderators tested explained the heterogeneity observed. Given the small number of studies, we may not have had adequate statistical power to detect moderating effects of the variables tested. Furthermore, there may have been other moderating variables that either were not tested in this study or measured in the primary reports, such as implementation fidelity, whether the intervention was theoretically informed, and so on, which could account for the differences in effects between studies. Because of the relatively small number of studies and the significant heterogeneity, caution must be used when interpreting and applying the findings of this review.

Court-based, school-based, and community-based interventions produced similar effects on attendance behaviors. The substantial similarity in mean effects across settings suggests that, when choosing to implement an intervention, one may choose from various settings and types of programs (school-, court-, or community-based). Given this finding, it seems reasonable for communities to select the setting and primary responsible organization based on ease of implementation, who has the most resources, or who is most invested in the program or outcomes. As there was significant heterogeneity within the groups of studies and few studies in some of the categories, it is important to note that there likely were not sufficient means to detect differences between interventions when there may, in fact, be real differences.

The focal modality utilized within the interventions—whether comprised primarily of a group, family, mentoring, or alternative education program or a contracting-only component/components—also produced statistically similar effects on attendance outcomes. Thus, there is no one modality that can be recommended over others. It is important to note that the within-group mean effects for the mentoring, alternative education, and contracting modalities were not statistically significantly different from zero. Due to the small sample size and the heterogeneity between studies, however, there likely was not sufficient power to detect group differences, especially since some groups only contained one or two studies within the group.

A key finding of this review and meta-analysis was the lack of available evidence to support the general belief that collaborative and multimodal interventions are more effective than simple, noncollaborative interventions. Although there is widespread support for, and many claims of greater effectiveness of, multimodal and/or collaborative programs (Bell, Rosen, & Dynlacht, 1994; Kearney, 2008; Kim & Streeter, 2006; Teasley, 2004; U.S. Department of Education, 1996), we did not find differences in mean effects between studies that utilized simple or noncollaborative interventions and those that were complex or collaborative. Although complex programs may have more intrinsic value and may be able to target several risk factors, potentially increasing their likelihood of success, implementation issues may reduce the potential effects of more complex programs. Single-modality interventions may be easier to implement and, therefore, more likely to be successful. More studies are needed to examine the effects of various interventions, including differential effects of different types of interventions in different settings that may account for why some collaborative interventions are successful while others are not.

Another important finding is the lack of overall clinical significance of interventions examined in the included studies. Although the effects of truancy interventions were positive and moderate, and the treatment group on average improved attendance by 4.69 days over the control group students, posttest attendance for most interventions remained at unacceptable levels. Even though students who receive an intervention do significantly better, as a whole, in their attendance than their control-group peers, many are still not achieving acceptable levels of attendance following the intervention.

The overall lack of reporting on, and statistical analysis of, demographic variables, particularly race/ethnicity and SES, was another surprising finding. Given that race and SES have been linked to absenteeism, the absence of the racial/ethnic and SES description of the participants was startling. In addition, the authors did not commonly utilize racial/ethnic or SES variables to compare treatment and comparison groups for equivalence nor look at possible differential effects of outcomes related to race or SES, both of which we would argue are imperative in research on outcomes of attendance interventions.

Implications for Practice and Policy

Due to the relatively small number of studies included in this synthesis, and the heterogeneous nature of the included studies, we believe that it is premature to recommend for or against the use of any of the interventions included in this analysis. That being said, these studies do represent the best empirical evidence currently available for outcomes of indicated programs targeting students with attendance problems. We believe that the findings from this review can provide some evidence and guidance, as well as some caution, for those who are concerned about, and trying to take action and develop policy to improve, attendance of truant students.

Overall, the findings from this study suggest that truant students benefit from interventions targeting attendance behaviors; thus, it is important and worthwhile to intervene with truant youth. Interventions that were implemented for only a couple of hours in duration and those implemented over the course of the school year produced substantially similar effects; thus it does not appear, at least in the short term, that the length of time for which the student receives the intervention either enhances or limits the effect on attendance. Because most studies did not assess outcomes beyond posttest, it is not known if, or for how long, the effects are sustained, or if longer-term interventions produce better outcomes over time.

The current literature espouses the use of collaborative and multimodal interventions. Interventions in this meta-analysis that were considered to be collaborative or multimodal produced mean effects that were substantially similar to those of simple interventions or those implemented by a single entity. This is encouraging in that it suggests that interventions may not need to be highly complex or involve multiple organizations or providers to have an impact on attendance outcomes. The evidence suggests that those who do not have significant resources or the time required to develop complex, collaborative programs can, nonetheless, make a difference and help truant students improve their attendance.

Although the interventions included in this study were, overall, found to improve attendance, the mean rates of absenteeism at posttest in most studies remained above acceptable levels. This finding indicates the need for additional work in developing more effective interventions and policies as well as in studying outcomes of interventions, particularly with vulnerable and at-risk populations.

The findings of this review have highlighted the lack of rigorous evidence to support many of the suggestions and recommendations being made by authors or program implementers. It seems that claims of success or effectiveness described in the literature and media are based on anecdotal evidence, or at best poorly executed evaluation studies, rather than on rigorous outcome research. Given this finding, it is important for practitioners and policy makers to be good consumers of evidence, rather than relying on anecdotal claims. Taking a “buyer beware” approach and being able to critically evaluate claims of effectiveness and research will be important to practitioners and policy makers who want to implement interventions that are based on rigorous evaluation and evidence.

In addition to becoming good consumers of evidence, it is very important for practitioners to be producers of evidence. There are many interventions throughout various countries that may be effective, but we cannot build the evidence base around what works to impact absenteeism if those interventions are not rigorously evaluated, reported, and disseminated. Those in the field doing the work of intervening with youth and families are well positioned to contribute to the evidence base, especially if they can carefully and thoroughly report what they are doing in their programs and use rigorous research design methods to examine outcomes.

Implications for Research

Despite the increased pressure for evidence-based practice and policy and the serious and widespread problem of truancy, there continues to be a paucity of research in the area of interventions to improve school attendance for chronic truant youth. Given the relatively small number of studies retrieved that met criteria for inclusion in this review, and the wide variety of interventions included in this review, it is obvious that there is a need for additional research in this area.

Although more research is needed, more of the same will not suffice. The studies included in this synthesis were plagued with methodological shortcomings, and a number of gaps in the evidence base were identified. Recommendations to improve the quality and fill gaps in the research are summarized in Table 11.

Table 11.

Summary of Recommendations to Improve Study Quality and Fill Research Gaps

Issue	Recommendation
Lack of Rigorous Study Designs	Utilize a comparison group design, preferably with random assignment
Missing Demographic Data	Provide adequate descriptions of the sample including: age, grade, race, SES, gender, special education status, and % attendance at baseline
Inadequate Descriptions of Intervention	Provide a detailed description of the intervention in such a way that the intervention could be replicated, including duration and intensity of sessions
Attrition	Keep attrition to a minimum. Clearly report attrition and reasons for lost cases
Small Sample Sizes	Keep sample size as large as feasible, taking into account issues of attrition, locating/enrolling participants, and student/family mobility
Inconsitency in Reporting and Measuring Attendance	Measure excused and unexcused absences and report separately. Report number of days absent/attended as well as the number of days possible or report the proportion of days attended/absent
Lack of Long-Term Follow-Up	Measure and report attendance at time points following the intervention, preferably a semester, school year, and beyond if possible
Inadequate Reporting of Data	Report the sample size, mean, and standard deviation for all outcomes measured, regardless of whether the results of statistical tests for that variable were significant
Lack of ethnic minority students	Additional studies are needed with students from various racial/ethnic backgrounds
Lack of court- and community-based interventions	Additional studies are needed to evaluate outcomes of court-based programs and clinic/community-based programs
Cost–benefit analysis	Data regarding the costs of the interventions and a cost–benefit analysis are needed in future studies
Missing information re implementation	Description and analysis of implementation issues is needed in future studies
Lack of consensus on definitions	Further research is needed to examine whether distinguishing students as school refusers as a distinct group is necessary
Few studies assessing long long-term outcomes	Studies need to examine longer-term outcomes related to attendance, dropout, and achievement

Note. SES = socioeconomic status.

Conclusions

There are hundreds of truancy interventions in operation with a goal of increasing attendance, many of which have been described in the literature as positively impacting the students and communities they are serving. Unfortunately, rigorous research to support truancy interventions is either not being conducted or is not being disseminated in a way that can inform others. Either way, evidence is not being built in a way that can add to the evidence base of effects of truancy interventions to inform practice and policy. In this era of evidence-based practice, No Child Left Behind, and numerous other initiatives at the local, state, and federal levels in which substantial amounts of money and efforts have been invested, it is surprising that the quantity and quality of outcome research of truancy is in such a paltry state.

In order to move the field forward, the various disciplines engaged in truancy research need to take a critical look at the barriers affecting research and dissemination. The social, political, and practical issues and barriers will need to be acknowledged, examined, and addressed if we hope to positively impact the attendance problem plaguing this country and others around the world.

Footnotes

Acknowledgments

We would like to thank Joshua Polanin and Ryan Williams for their assistance in coding studies for this review and the NCSE for sharing their database of registered programs with us.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship and/or publication of this article: The authors would like to acknowledge the Campbell Collaboration, the Arthur J. Schmitt Foundation, the Meadows Center for Preventing Educational Risk, and the Institute of Education Sciences grant R324B080008 for providing support for this study. The manuscript content does not necessarily represent the positions or policies of the funding agencies.

References

Attwood

Croll

(2006). Truancy in secondary school pupils: Prevalence, trajectories and pupil perspectives. Research Papers in Education, 21, 467–484.

Baker

M. L.

Sigmon

J. N.

Nugent

M. E.

(2001). Truancy reduction: Keeping students in school. Bulletin of the Office of Juvenile Justice and Delinquency Prevention, September, 1–15.

Bell

A. J.

Rosen

L. A.

Dynlacht

(1994). Truancy intervention. The Journal of Research and Development in Education, 27, 203–211.

Borenstein

Hedges

Higgins

Rothstein

(2005). Comprehensive meta-analysis (Version 2). Englewood, NJ: Biostat.

Bowen

N. K.

Bowen

G. L.

Ware

W. B.

(2002). Neighborhood social disorganization, families, and the educational behavior of adolescents. Journal of Adolescent Research, 17, 468–490.

*Converse

Lignugaris/Kraft

(2009). Evaluation of a school-based mentoring program for at-risk middle school youth. Remedial and Special Education, 30, 33–46.

Cooper

(1998). Synthesizing research: A guide for literature reviews (3rd ed.). Thousand Oaks, CA: Sage.

Corville-Smith

Ryan

B. A.

Adams

G. R.

Dalicandro

(1998). Distinguishing absentee students from regular attenders: The combined influence of personal, family, and school factors. Journal of Youth and Adolescence, 27, 629–640.

Davies

J. D.

Lee

(2006). To attend or not to attend? Why some students choose school and others reject it. Support for Learning, 21, 204–209.

10.

*DeSocio

VanCura

Nelson

L. A.

Hewitt

Kitzman

Cole

(2007). Engaging truant adolescents: Results from a multifaceted intervention pilot. Preventing School Failure, 51, 3–11.

11.

*Fantuzzo

Grimm

Hazen

(2005). Project Start: An evaluation of a community-wide school-based intervention to reduce truancy. Psychology in the Schools, 42, 657–667.

12.

*Flanagan

(2006). The effects of a positive behavior support program on the attendance behaviors of students with and without disabilities in a large inner-city public high school. Dissertation Abstracts International, 67, (07) (UMI No. 3225140).

13.

Garry

E. M.

(1996). Truancy: First step to a lifetime of problems. OJJDP: Juvenile Justice Bulletin, October, 1–7.

14.

Glass

G. V.

McGaw

Smith

M. L.

(1981). Meta-analysis in social research. Thousand Oaks, CA: Sage.

15.

*Glover

O. M.

(1990). Effectiveness of group intervention and peer support on truancy. Masters Abstracts International, 29, 50 (UMI No. 1340763).

16.

Goldstein

J. S.

Little

S. G.

Akin-Little

(2003). Absenteeism: A review of the literature and school psychology’s role. The California School Psychologist, 8, 127–139.

17.

Hedges

L. V.

(1981). Distribution theory for Glass’s estimator of effect size and related estimators. Journal of Educational Statistics, 6, 107–128.

18.

Hedges

L. V.

Olkin

(1985). Statistical methods for meta-analysis. Orlando, FL: Academic Press.

19.

*Herrick

(1992). Effects of an attendance incentive program for chronically absent elementary school students. Dissertation Abstracts International, 54, 2010A (UMI No. 933147).

20.

*Hess

A. M.

(1990). The effects of contingency contracting and parent training on the truant behavior and classroom performance of students with handicaps. Dissertation Abstracts International, 51, 1576A (UMI No. 9030200).

21.

*Hess

A. M.

Rosenberg

M. S.

Levy

G. K.

(1990). Reducing truancy in students with mild handicaps. Remedial and Special Education, 11, 14–19.

22.

*Hubin

M. L.

(2000). Stop Truancy Project (SToP): A program evaluation. Unpublished MSW clinical research paper, College of St. Catherine and University of St. Thomas, St. Paul, MN.

23.

*Johncox

J. F.

(1994). Intervention effects for treating student nonattenders and their families. Unpublished MSW clinical research paper, College of St. Catherine & University of St. Thomas, St. Paul, MN.

24.

*Johnson

Syropoulos

(1996). Evaluation report: High school intervention centers program 1995-1996. Detroit, MI: Detroit Public Schools Office of Research, Evaluation, and Testing (ERIC Document Reproduction Service No. ED 407474).

25.

Kearney

C. A.

(2008). School absenteeism and school refusal behavior in youth: A contemporary review. Clinical Psychology Review, 28, 451–471.

26.

Kim

J. S.

Streeter

C. L.

(2006). Increasing school attendance: Effective strategies and interventions. In Franklin

Harris

M. B.

Allen-Meares

(Eds.), The school services sourcebook: A guide for school-based professionals (pp. 397–404). New York, NY: Oxford University Press.

27.

Klima

Miller

Nunlist

(2009). What works? Targeted truancy and dropout programs in middle and high school (Olympia: Washington State Institute for Public Policy, No. 09-06-2201). Retrieved from http://www.wsipp.wa.gov/pub.asp?docid=09-06-2201

28.

Lauchlan

(2003). Responding to chronic non-attendance: A review of intervention approaches. Educational Psychology in Practice, 19, 133–146.

29.

*Lawson

C. L.

(1990). Peer tutoring as a factor in reducing school truancy and increasing performance among African-American elementary school students in Chicago. Dissertation Abstracts International, 52, 1710A (UMI No. 9129846).

30.

Lipsey

M. W.

Wilson

D. B.

(2001). Practical meta-analysis. Thousand Oaks, CA: Sage.

31.

Malcolm

Wilson

Davidson

Kirk

(2003). Absence from school: A study of its causes and effects in seven LEAs (Department for Education and Skills Research Report RR424). Sherwood Park, England: National Foundation for Educational Research. Retrieved from http://nces.ed.gov/programs/coe/indicator_mis.asp

32.

Maynard

B. R.

McCrea

K. T.

Pigott

T. D.

Kelly

M. S.

(2009). Interventions intended to increase school attendance in students attending primary or secondary schools: Review protocol. Campbell Systematic Reviews. Retrieved from www.campbellcollaboration.org

33.

Maynard

B. R.

McCrea

K. T.

Pigott

T. D.

Kelly

M. S.

(2012). Indicated truancy interventions: Effects on school attendance among chronic truant students. Campbell Systematic Reviews. Retrieved from www.campbellcollaboration.org. doi: 10.4073/csr.2012.10

34.

Maynard

B. R.

Salas-Wright

C. P.

Vaughn

M. G.

Peters

K. E.

(2012). Who are truant youth? Examining distinctive profiles of truant youth using latent profile analysis. Journal of Youth and Adolescence. Advance online publication. doi: 10.1007/s10964-012-9788-1

35.

National Center for Education Statistics. (2006). The condition of education 2006. Washington, DC: US Department of Education. Retrieved October 10, 2008, from http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2006071

36.

National Center for School Engagement (NCSE). (2007). Pieces of the truancy jigsaw: A literature review. Retrieved October 3, 2007, from www.schoolengagement.org

37.

Petrides

K. V.

Chamorro-Premuzic

Frederickson

Furnham

(2005). Explaining individual differences in scholastic behavior and achievement. British Journal of Educational Psychology, 75, 239–255.

38.

Reid

(1999). Truancy and schools. London, England: Routledge.

39.

Reid

(2002). Truancy: Short and long-term solutions. London, England: RoutledgeFalmer.

40.

Reimer

M. S.

Dimock

(2005). Best practices and model truancy programs. Clemson, SC: National Dropout Prevention Center/Network.

41.

*Seamans

C. L.

(1996). Brief family systems intervention with inner-city truant youths. Masters Abstracts International, 35, 349 (UMI No. 1381227).

42.

*Sherriff

I. H.

(1990). A multi-disciplinary approach to the management of non-school attendance (Doctoral dissertation, University of Leicester, 1990). Dissertation Abstracts International, 52, 2415A.

43.

Stahl

A. L.

(2008). Petitioned status offense cases in juvenile courts, 2004. OJJDP Fact Sheet, 2, 1–2.

44.

Sutphen

R. D.

Ford

J. P.

Flaherty

(2010). Truancy interventions: A review of the research literature. Research on Social Work Practice, 20, 161–171.

45.

Teasley

M. L.

(2004). Absenteeism and truancy: Risk, protection, and best practice implications for social workers. Children and Schools, 26, 117–128.

46.

*Tichenor

R. M. S.

(1991). Making it in middle school: An evaluation of the effects of a guidance program on the school attendance patterns of at-risk sixth-graders. Dissertation Abstracts International, 52, 3831A (UMI No. 9211808).

47.

U.S. Department of Education. (1996). Manual to combat truancy. Washington, DC: Author.

48.

Valentine

J. C.

Pigott

T. D.

Rothstein

H. R.

(2009). How many studies do you need? A primer on statistical power for meta-analysis. Journal of Education and Behavioral Statistics, 35, 215–247.