CoChat
Sign Up
Cited by 636Early Childhood Education and DevelopmentYouth Development and Social SupportChild and Adolescent Psychosocial and Emotional Development

The state of evidence for social and emotional learning: A contemporary meta-analysis of universal school-based SEL interventions

Open access
Cipriano, Christina, Strambler, Michael J, Naples, Lauren H, Ha, Cheyeon, Kirk, Megan, Wood, Miranda, Sehgal, Kaveri, Zieher, Almut K +6 more·2023·Child Development
DOI 10.1111/cdev.13968
Open this paper in CoChat — free accountOpen PDF

Abstract

This article provides a systematic review and meta-analysis of the current evidence for universal school-based (USB) social and emotional learning (SEL) interventions for students in kindergarten through 12th grade available from 2008 through 2020. The sample includes 424 studies from 53 countries, reflecting 252 discrete USB SEL interventions, involving 575,361 students. Results endorsed that, compared to control conditions, students who participate in USB SEL interventions experienced significantly improved skills, attitudes, behaviors, school climate and safety, peer relationships, school functioning, and academic achievement. Significant heterogeneity in USB SEL content, intervention features, context, and implementation quality moderated student experiences and outcomes. Strengths and limitations of this evidence and implications for future USB SEL research, policy, and practice are discussed.

Integrity signals

Registry

Crossref · 10.1111/cdev.13968

Cites

111 of 111 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

9 of 53 cited sources checked · 4 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    There is a lack of evidence in the effectiveness of Diversity, Equity, and Inclusion initiatives (DEI).

    “especially given the lack of evidence in the effectiveness of Diversity, Equity, and Inclusion initiatives (DEI; Paluck et al., 2021;Forscher et al., 2019;Lai et al., 2016).”
    Quote in the paperSubstantial claim3 of 3 sources checked
    • cites Paluck et al., 2021 source

      The Applied Interventions section says there is no broad evidence base for judging the efficacy of diversity training, and the Outcome Measurement section describes evidence for implicit-prejudice reduction as thin. Those findings support concern about particular DEI-related trainings, but they do not establish a lack of evidence for the effectiveness of DEI initiatives as a whole. Page numbers were not supplied with the passages. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites Forscher et al., 2019 source

      The Conclusion reports little evidence that changing implicit measures changes behavior, including in intergroup-bias interventions. But the paper assesses procedures aimed at implicit measures, not the effectiveness of DEI initiatives generally; the claim extends its finding beyond the interventions studied. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites Lai et al., 2016 source

      The General Discussion (p. 22) reports that nine interventions reduced implicit racial preferences immediately, but none remained effective after a delay. This supports concern about lasting effects of the tested bias-reduction interventions, not a general claim that evidence is lacking for the effectiveness of DEI initiatives as a whole. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
  • Findingown data

    Compared to control conditions, students who participate in USB SEL interventions experienced significantly improved peer relationships.

    “Results endorsed that, compared to control conditions, students who participate in USB SEL interventions experienced significantly improved skills, attitudes, behaviors, school climate and safety, peer relationships, school functioning, and academic achievement.”
    Quote in the paperSubstantial claim
  • Findingown data

    Compared to control conditions, students who participate in universal school-based (USB) social and emotional learning (SEL) interventions experienced significantly improved skills.

    “Results endorsed that, compared to control conditions, students who participate in USB SEL interventions experienced significantly improved skills, attitudes, behaviors, school climate and safety, peer relationships, school functioning, and academic achievement.”
    Quote in the paperSubstantial claim
  • Findingown data

    Compared to control conditions, students who participate in USB SEL interventions experienced significantly improved school functioning.

    “Results endorsed that, compared to control conditions, students who participate in USB SEL interventions experienced significantly improved skills, attitudes, behaviors, school climate and safety, peer relationships, school functioning, and academic achievement.”
    Quote in the paperSubstantial claim
  • Findingown data

    Significant heterogeneity in USB SEL content, intervention features, context, and implementation quality moderated student experiences and outcomes.

    “Significant heterogeneity in USB SEL content, intervention features, context, and implementation quality moderated student experiences and outcomes.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. 29
  • May combine two claims. 27

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    The review adhered to a predetermined, peer-reviewed protocol to strengthen transparency through pre-registration.

    “This systematic review and meta-analysis followed the contemporary PRISMA guidelines (Page et al., 2021) and adhered to a predetermined, peer-reviewed protocol to strengthen transparency through pre-registration (Nosek et al., 2012).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Nosek et al., 2012 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    This systematic review and meta-analysis followed the contemporary PRISMA guidelines.

    “This systematic review and meta-analysis followed the contemporary PRISMA guidelines (Page et al., 2021) and adhered to a predetermined, peer-reviewed protocol to strengthen transparency through pre-registration (Nosek et al., 2012).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Page et al., 2021 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    The search strategy was peer reviewed using the Peer Review of Electronic Search Strategies (PRESS) Checklist.

    “The search strategy was peer reviewed using the Peer Review of Electronic Search Strategies (PRESS) Checklist (McGowan et al., 2016).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites McGowan et al., 2016 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    The study followed the PRISMA framework and methodological guidance for research in education and psychology.

    “In sum, we followed the PRISMA framework and methodological guidance for research in education and psychology to report findings and reach conclusions (Page et al., 2021;Pigott & Polanin, 2020;Polanin et al., 2020).”
    Quote in the paperNot a substantial claim1 of 3 sources checked
    • cites Page et al., 2021 source
    • cites Pigott & Polanin, 2020 source
    • cites Polanin et al., 2020 source

  • May combine two claims.

    Study design and conductcitation

    Authors reached 'almost perfect' interrater reliability for both stages of screening (0.93%-1% for title and abstract, and 0.81%-0.98% agreement for full text).

    “Authors reached”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites McHugh, 2012 source

      The section ‘Theoretical issues in measurement of rater reliability’ reports that kappa values of 0.81–1.00 have been described as ‘almost perfect’ agreement. That classification concerns kappa, whereas the claim describes percentage agreement; the paper does not report the citing authors’ screening-stage results. (LLM, conf=0.83)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductcitation

    We double coded each indicator and merged codes into a composite index score for analyses.

    “We double coded each indicator and merged codes into a composite index score for analyses (Kremer et al., 2015;Lester et al., 2020).”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Kremer et al., 2015 source

      The Study Selection and Data Extraction section reports that two authors independently coded moderator and effect-size data, but descriptive data were mostly coded by one author. The Assessing Risk of Bias section says ratings were made separately for each domain, and Statistical Procedures describes analyzing effect sizes without averaging them into a composite index. The source supports double coding of some data, not double coding every indicator or merging codes into a composite index score. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites Lester et al., 2020 source

  • May combine two claims.

    Findingown data

    The meta-analysis sample includes 424 studies from 53 countries, reflecting 252 discrete USB SEL interventions, involving 575,361 students.

    “The sample includes 424 studies from 53 countries, reflecting 252 discrete USB SEL interventions, involving 575,361 students.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    We identified 252 different SEL interventions, 210 named and 42 unnamed.

    “We identified 252 different SEL interventions, 210 named and 42 unnamed, in our review.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    For six outcome domains (50%) follow-up effect sizes remained significant and yielded a small to small-moderate effect in favor of the SEL intervention condition.

    “For six outcome domains (50%) followup effect sizes remained significant and yielded a small to small-moderate effect in favor of the SEL intervention condition.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: The mean correlation within and across studies is assumed to be ρ = .60.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Most studies used an education as usual or waitlist control (381; 72.9%).

    “most studies used an education as usual or waitlist control (381; 72.9%).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    No evidence of significant mean differences on discipline, physical health, nor family relationships between intervention and control groups at postintervention (p > .05).

    “We found no evidence of significant mean differences on discipline, physical health, nor family relationships between the intervention and control groups at postintervention (p > .05).”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: The mean correlation within and across studies is assumed to be ρ = .60.

  • May combine two claims.

    Findingown data

    Among all relationships, only early adolescents trended toward marginally improved civic attitudes and behaviors (g = 0.371) compared to adolescents (p = 0.07).

    “among all relationships only early adolescents trended toward marginally improved civic attitudes and behaviors (g = 0.371) compared to adolescents (p = 0.07).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Studies were roughly split between randomized control trials (237; 55.9%) and quasi-experimental designs (187; 44.1%).

    “Studies were roughly split between randomized control trials (237; 55.9%) and quasi-experimental designs (187; 44.1%)”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Only 44.6% of studies reported any implementation data, with 21% of those reporting the level of implementation as Moderate or High.

    “Only 44.6% reported any implementation data, with 21% of those reporting the level of implementation as Moderate or High.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    The studies analyzed reflect 2049 SEL outcomes.

    “The studies analyzed reflect 2049 SEL outcomes.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Domain 1 (SEL Skills) includes 119 different measures reflecting 20 constructs and 30 outcomes that combined 2 or more constructs.

    “Domain 1 (SEL Skills) includes 119 different measures reflecting 20 constructs and 30 outcomes that combined 2 or more constructs.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Published studies had an average effect size of g = 0.220 (95% CI [0.191, 0.256]) whereas unpublished studies had g = 0.164 (95% CI [0.096, 0.231]), indicating a meaningful difference between these two classes of studies.

    “On average, published studies had an average ES of g = 0.220 (95% CI [0.191, 0.256]) whereas unpublished was g = 0.164 (95% CI [0.096, 0.231]) indicating a meaningful difference between these two classes of studies.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The review did not find evidence of effects for family relationships and physical health outcomes.

    “We did not find evidence of effects for two domains, family relationships and physical health outcomes.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    The SDQ is reflected 154 times across 7 outcome domains in the study.

    “The SDQ is reflected 154 times across 7 outcome domains in our study.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingnone given

    The same measures were operationalized as different constructs across and within the twelve outcome domains examined, and they were inconsistently presented and justified.

    “The same measures were operationalized as different constructs across and within the twelve outcome domains examined, and they were inconsistently presented and justified.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Egger's regression test indicated a potential for publication bias in the outcome domains of SEL skills (t[320] = 4.10, p < .001), prosocial behavior (t[197] = 3.30, p = .001), and school climate (t[114] = 4.13, p < .001).

    “and in the outcome domains of SEL skills (t[320] = 4.10, p < .001),prosocial behavior (t[197] = 3.30, p = .001),and school climate (t[114] = 4.13, p < .001).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    The most frequently reported assessment in the review was the Strengths and Difficulties Questionnaire (SDQ).

    “The most frequently reported assessment in our review was the Strengths and Difficulties Questionnaire (SDQ; Goodman, 2001)”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Goodman, 2001 source

  • May combine two claims.

    Findingown data

    All studies that reported participants' gender did so in binary terms, obscuring the identity of participants identifying as non-binary or transgender/transsexual.

    “all studies that reported participants' gender did so in binary terms, obscuring the identity of participants identifying as non-binary or transgender/transsexual.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingreasoning

    Studies guided by a mindfulness framework were common in the literature, but these studies differed based on the type of mindfulness framework used.

    “studies guided by a mindfulness framework were common in the literature, but these studies differed based on the type of mindfulness framework used”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In the review, student grades represented only one third of the academic indicators, with standardized tests accounting for two thirds of the distribution of scores included in analysis.

    “In our review, student grades only represented one third of the academic indicators, with standardized tests accounting for two thirds of the distribution of scores included in analysis (See Table S10).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Durlak et al. (2011) found that programs using all four SAFE components (Sequenced, Active, Focused, Explicit) were more effective than those that did not include these components.

    “Durlak et al. (2011) found that programs that”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Durlak et al., 2011 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Durlak et al. (2011) found that programs that monitored implementation (i.e., fidelity) were more effective than those that did not.

    “Durlak et al. (2011) found that programs that monitored implementation (i.e., fidelity) were more effective than those that did not.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Durlak et al., 2011 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    Articles were double screened: first title and abstracts, then full-text.

    “Articles were double screened: first title and abstracts, then full-text.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Conflicts were resolved through discussion and consensus.

    “Conflicts were resolved through discussion and consensus.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    Two authors independently reviewed all excluded studies and resolved any conflicts by discussion.

    “To ensure reliability on study exclusion, two authors independently reviewed all excluded studies and resolved any conflicts by discussion.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The analytic files are available on Open Science Framework (OSF) alongside supplemental papers, data, instructions, and code for analysis to support reproducibility.

    “The analytic files reflected in this paper are available on Open Science Framework (OSF) alongside two supplemental papers, data, instructions, and code for analysis to support reproducibility.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    Raters' agreements on continuous variables were strong, all above 0.90, and agreement on categorical variables ranged from .78-.92.

    “Raters' agreements on continuous variables were strong, all above 0.90, and agreement on categorical variables ranged from .78-.92, respectively.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    We assessed risk of bias using the adapted Joanna Briggs Institute Critical Appraisal tool (2020) for quasi-experimental and randomized controlled trials.

    “Study design variables included report format, and an examination of risk of bias using the adapted Joanna Briggs Institute Critical Appraisal tool (2020) for quasiexperimental and randomized controlled trials.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Joanna Briggs Institute, 2020

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    Overall effect sizes for published studies were compared to unpublished studies.

    “we compared overall effect sizes for published to unpublished studies.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    CHE models with RVE were used to estimate mean effects, accounting for clustering and correlated effect sizes.

    “To estimate mean effects, we conducted Correlated and Hierarchical Effects (CHE) models (Pustejovsky & Tipton, 2022) with Robust Variance Estimation (RVE).”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: The mean correlation within and across studies is assumed to be ρ = .60.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    An a priori power analysis was conducted using conservative values from Durlak et al. (2011).

    “An a priori power analysis with a calculator (Quintana & Tiebel, 2019) was conducted based on equations from Valentine et al. (2010) using conservative values from Durlak et al. (2011)”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Quintana & Tiebel, 2019

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Valentine et al., 2010 source
    • cites Durlak et al., 2011 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    Publication bias was assessed through funnel plots, contour-enhanced funnel plots, and Egger's regression tests accounting for nested data.

    “we assessed the extent publication bias may have impacted findings through reviewing funnel plots, contour-enhanced funnel plots, and Egger's regression tests where we accounted for the nested structure of the data (Egger et al., 1997).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Egger et al., 1997 source

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    Whenever possible, adjusted post and follow-up effect sizes were calculated using pre-intervention differences between intervention and control groups.

    “Whenever possible, we calculated adjusted post and follow-up effect sizes using any pre-intervention differences that occurred between intervention and control groups.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    Power was 0.63 for N = 7, 0.78 for N = 10, and 0.91 for N = 15.

    “With these conservative assumptions, power was 0.63 for N = 7, 0.78 for N = 10, and 0.91 for N = 15.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: The power analyses assume high heterogeneity of I2 = 0.75.

  • May combine two claims.

    Study design and conductown data

    The meta-analytic sample did not differ from the full sample in general study features, sample ecology, or USB SEL intervention features.

    “The meta-analytic sample did not differ from the full sample in general study features, sample ecology, or USB SEL intervention features (See Table S5 for comparison).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductreasoning

    All studies were double coded and more than half were triple coded to estimate coding reliability.

    “All studies were double coded and due to wide disparities in SEL science and reporting, more than half of the studies were triple coded to estimate coding reliability.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning

    Effect size outliers more than 2 SD from the grand mean were reset to exactly the 2 SD points to guard against them unduly affecting results.

    “We reset these outliers to be exactly at the 2 SD points to guard against them unduly affecting results.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation

    The finding that published studies had larger average effects than unpublished studies is consistent with other meta-review research that found the same trend.

    “This finding is consistent with other meta-review research that found the same trend (Chow & Ekholm, 2018).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Chow & Ekholm, 2018 source

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Premisecitation

    Simply knowing one's identity says little about how central that identity is to oneself versus other aspects of one's identity.

    “Simply knowing one's identity says little about how central that identity is to oneself versus other aspects of one's identity; the experience of identity is personal and varied, and intersects with how students interact with and develop skills from SEL programming (Cipriano et al., 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Cipriano et al., 2023 source

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

    Interpretationnone given

    The amassed data of the report only reflects the beginning of the contemporary evidence suggesting SEL interventions support healthy and inclusive school climates, increased school achievement, and healthy attitudes and behaviors among students.

    “The amassed data of this report only reflects the beginning of the contemporary evidence suggesting SEL interventions support healthy and inclusive school climates, increased school achievement, and healthy attitudes and behaviors among students.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Students who participate in SEL programs in their classrooms and schools are more engaged and have better quality relationships with their teachers.

    “students who participate in SEL programs in their classrooms and schools are more engaged and have better quality relationships with their teachers (Brown et al., 2010;Hagelskamp et al., 2013).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Brown et al., 2010 source
    • cites Hagelskamp et al., 2013 source

  • May combine two claims.

    Interpretationnone given

    SEL interventions vary widely and are not one cohesive approach.

    “SEL interventions vary widely and are not one cohesive approach.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Wolf and Harbatkin (2023) used What Works Clearinghouse data to categorize outcome measures and documented how average effect sizes varied for outcome measure type by whether the group who developed the measure potentially had a stake in the intervention or not.

    “Wolf and Harbatkin (2023) used data from the What Works Clearinghouse to categorize outcome measures in educational interventions and documented how average effect sizes varied for outcome measure type by whether the group who developed the measure potentially had a stake in the intervention (non-independent) or not (independent).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Wolf & Harbatkin, 2023 source

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The high missingness of identifiers, alongside significant heterogeneity in identity descriptors, would result in biased estimates of effects if aggregated as proposed.

    “the high missingness of identifiers, alongside significant heterogeneity in identity descriptors would result in biased estimates of effects if aggregated as proposed.”
    Quote in the paperSubstantial claim
49 of 49

  • The authors state this as an assumption.

    The mean correlation within and across studies is assumed to be ρ = .60.

    “To be conservative, we assigned a high value of ⍴ = .60for post and follow-up metaanalyses an assumed correlation within and between studies.”

  • The authors state this as an assumption.

    The power analyses assume high heterogeneity of I2 = 0.75.

    “All three power analyses assume high heterogeneity of I 2 = 0.75.”

  • The paper relies on this without support.

    The effect sizes from the included studies are sufficiently comparable to be pooled into overall domain-level estimates despite heterogeneity in intervention content, implementation, and outcomes.

    “Our comprehensive search strategy and expansive inclusion criteria resulted in a highly heterogeneous sample in terms of SEL intervention content, implementation, and outcomes.”
    “We found a significant overall effect of USB SEL interventions on student outcomes despite this variability, suggesting the robustness of findings.”
3 of 3

Statistics recomputed

Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.

  • ConsistentQ(1861) = 27101.93reported p < .001 · recomputed p < .001

    “Q (1861) = 27,101.93,p < .001”

  • ConsistentQ(326) = 3293reported p < .001 · recomputed p < .001

    “Q (326) = 3293.00,p < .001”

2 of 2

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Discussion effect sizes differ from the reported results

    Major

    The discussion gives effect sizes of 0.228, 0.193, and 0.149 for SEL skills, prosocial behaviors, and emotional distress. The results give 0.219, 0.178, and 0.140 for those same domains.

  • Power-analysis sample size changes from 12 to 15

    Minor

    The methods say the three power analyses used N = 7, 10, and 12, but the subsequent results give power for N = 7, 10, and 15.

  • Percentage of named interventions uses the wrong denominator

    Minor

    Of 252 interventions, 210 named interventions constitute 83.3%, not the table's 90.1%.

    Reported 90.1 · recomputed 83.333333

  • Child-reported outcome share does not match the stated total

    Minor

    Using the stated total of 2,049 outcomes, 1,185 child-reported outcomes account for 57.8%, not 53.5%.

    Reported 53.5 · recomputed 57.833089

  • Perspective-skill study count differs between table and prose

    Minor

    Table 2 lists 128 studies for perspectives, but the descriptive text lists 125 for the same category.

  • Family-engagement prose overlooks the largest listed category

    Minor

    The prose calls training and materials at home the most frequent family-engagement modes, with counts of 29 and 21. Table 2 lists letters or emails sent home for 175 studies, more than either mode.

  • Academic-achievement effect size is printed as 112

    Minor

    The results narrative reports an academic-achievement Hedges' g of 112, while the table note reports 0.112 for that outcome.

  • Control-condition percentage does not match the study count

    Minor

    The stated 381 studies represent 89.9% of the 424 reviewed studies, not 72.9%.

    Reported 72.9 · recomputed 89.858491

  • Follow-up results name five significant domains, not six

    Minor

    The follow-up section states that six domains had significant effects, but its subsequent list names three improved domains and two with reductions. It then describes two further domains as only marginally significant.

9 of 9

Cites

Flagged first, then by citations

Research works printed but not linked: 30

  1. Berg, 2017, The intersection of school climate and social and emotional development
  2. Berg, 2017, Identifying, defining, and measuring social and emotional competencies
  3. Center for Applied Special Technology [CAST], 2018, Universal design for learning guidelines version 2.2
  4. Centers for Disease Control and Prevention [CDC], 2019, Youth risk behavior survey: Data summary and trends report 2009–2019
  5. Cipriano, Methodological Guidance Paper: Conducting a High Quality Evaluation of an SEL Intervention
  6. Collaborative for Academic, Social, and Emotional Learning, 2013, Effective social and emotional learning programs: Preschool and elementary school edition
  7. Collaborative for Academic, Social, and Emotional Learning, 2015, Effective social and emotional learning programs: Middle and high school edition
  8. Collaborative for Academic, Social and Emotional Learning [CASEL], 2017, Sample teaching strategies to support core competencies of social and emotional learning
  9. Collaborative for Academic, Social and Emotional Learning [CASEL], 2020, What is SEL?
  10. Congress, 2020, Coronavirus Aid, Relief, and Economic Security Act (CARES Act). P.L., 116-136
  11. Dusenbury, 2019, An examination of frameworks for social and emotional learning (SEL) reflected in State K-12 learning standards
  12. Dymnicki, 2020, Advancing the use of core components of effective programs: Suggestions for researchers publishing evaluation results
  13. Ecological Approaches to Social Emotional Learning [EASEL], Explore EASEL: Identify related skills across frameworks
  14. Elias, 2019, (Eds.), Keeping students safe and helping them thrive: A collaborative handbook on school safety, mental health, and wellness, 185
  15. Every Student Succeeds Act of 2015, 2015, Pub. L. No. 114-95 § 114 Stat. 1177
  16. Joanna Briggs Institute, 2020, Checklist for systematic reviews and research syntheses
  17. Jones, The Taxonomy Project
  18. Jones, 2016, What is the same and what is different?
  19. Jones, 2017, Navigating SEL from the inside out: Looking inside & across 25 leading SEL programs—A practical resource for schools and OST providers
  20. Krachman, 2017, The scale of our investment in social and emotional learning
  21. Organization for Economic Cooperation and Development, 2017, Social and emotional skills well-being, connectedness and success
  22. Osher, 2020, Thriving, robust equity, and transformative learning & development: A more powerful conceptualization of the contributors to youth success
  23. Pustejovsky, 2022, clubSandwich: Cluster-Robust (sandwich) variance estimators with small-sample corrections
  24. RAND, 2020, RAND Education Assessment Finder
  25. Rimm-Kaufman, 2015, Handbook of social and emotional learning: Research and practice, 151
  26. Strambler, 2023, Social and emotional learning in action: Creating systemic change in schools, 281
  27. TransformEd, 2020, Trauma-informed SEL toolkit
  28. Wanless, 2020, Why social justice and equity is essential to SEL
  29. What Works Clearinghouse, 2017, Procedures handbook (Version 4.0)
  30. World Health Organization, 2003, Skills for health: Skills-based health education including life skills: An important component of a child-friendly/health-promoting school

Cited by

Sorted by citations

Related

Authors
Cipriano, C., Strambler, M. J., Naples, L. H., Ha, C., Kirk, M., Wood, M., Sehgal, K., Zieher, A. K., Eveleigh, A., McCarthy, M., Funaro, M., Ponnock, A., Chow, J. C., & Durlak, J.
Affiliations
  • Yale University (US)
  • University of Rochester (US)
  • The Child Study Center, Yale School of Medicine , New Haven, Connecticut ,
  • University of Maryland, College Park (US)
  • Loyola University Chicago (US)
Corresponding author
Cipriano, Christina
ORCID
Venue
Child Development
Publisher
Oxford University Press (OUP)
Published
2023-09-01
Volume
94
Issue
5
Pages
1181-1204
ISSN
  • 0009-3920 (ISSN-L)
  • 1467-8624
DOI
10.1111/cdev.13968
OpenAlex
W4384277430
Cited by
636
References
109
Citations by year
2026: 300 · 2025: 228 · 2024: 95 · 2023: 12 · 2014: 1
FWCI
38.49
Type
journal article
Language
en
Open access
Licence
Locations
  • Child Development (publishedVersion)
  • PubMed (publishedVersion)
Funders
Oak Foundation (OCAY-19-407, OFIL-22-040, OCAY‐19‐407, OFIL‐22‐040)
Topics
  • Social Sciences › Education › Early Childhood Education and Development
  • Social Sciences › Safety Research › Youth Development and Social Support
  • Social Sciences › Psychology › Clinical Psychology › Child and Adolescent Psychosocial and Emotional Development
Keywords
social-emotional learning, universal school-based intervention, academic performance, school climate, student behavior, peer relationships, school safety, implementation quality
MeSH terms
Social Learning, Attitude, Emotions, Humans, Learning, Schools

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-11.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-11.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review