The DOI matches its Crossref record; 97 of the 101 references it cites were found in a registry or an index; 5 of the 23 sources its claims cite were checked for support.
Concept mapping in STEM education: a meta-analysis of its impact on students' achievement (2004–2023)
Open accessAbstract
Abstract Concept mapping is a powerful educational tool that facilitates students' organization and integration of knowledge, leading to improved understanding and retention. Numerous studies have highlighted the potential of concept mapping to enhance students' achievement. However, comprehensive quantitative analyses focusing on its impact specifically in STEM education are limited. This meta-analysis synthesizes findings from 37 studies conducted between 2004 and 2023, identifying a moderate overall effect size of concept mapping on students' achievement in STEM disciplines (ES = 0.630). The analysis of six moderating variables (subject, educational level, duration of the experiment, concept map construction mode, learning equipment, and pedagogical approach) reveals that concept mapping is particularly effective in science and technology, with the greatest impact observed at the secondary education level. Interventions lasting between one week and one month yielded the most significant improvements in learning outcomes, especially when students independently constructed their concept maps. Both digital and paper-based tools demonstrated moderate and comparable effects on students' achievement. In addition, among various pedagogical approaches, contextual and inquiry-based pedagogical approaches are more effective when integrated into STEM education. This study provides recommendations for future research and practical implementation of concept mapping in educational contexts.
Overview
CoChat Summary
TL;DR
This meta-analysis of 35 papers yielding 37 effect sizes (2004–2023) found a moderate overall effect of concept mapping on STEM achievement (Hedges’ g = 0.630, 95% CI [0.469, 0.790], p < 0.001). Moderator analyses showed no statistically significant between-group differences, but numerically strongest effects appeared for science/technology, secondary education, 1 week–1 month interventions, and student-constructed maps.
Key findings
- Overall effect was moderate: g = 0.630 (95% CI [0.469–0.790], p < 0.001), with high heterogeneity (I² = 76.145, Q = 150.909, p < 0.001); a random-effects model was used.
- Subject differences were not statistically significant (Q = 3.145, p = 0.790); physics had the largest science effect (g = 0.782, p < 0.001), followed by earth science (g = 0.679, p < 0.001), biology (g = 0.632, p < 0.001), chemistry (g = 0.514, p = 0.138), technology (g = 0.769, p = 0.029), engineering (g = 0.363, p = 0.291), and mathematics (g = 0.217, p = 0.512).
- Educational level differences were not statistically significant (Q = 1.232, p = 0.745); upper secondary had g = 0.780 (p = 0.001), lower secondary g = 0.755 (p < 0.001), primary g = 0.576 (p < 0.001), and higher education g = 0.552 (p = 0.001).
- Duration, construction mode, equipment, and pedagogy showed no significant between-group differences (Q = 1.624, p = 0.804; Q = 3.635, p = 0.304; Q = 0.107, p = 0.948; Q = 1.837, p = 0.871), but numerically highest effects were 1 week–1 month (g = 0.768, p < 0.001), student-constructed maps (g = 0.933, p < 0.001), mobile devices (g = 0.687, p < 0.001), and contextual learning (g = 0.873, p = 0.020).
Methods
- PRISMA-guided search of six databases (Web of Science, ERIC, ProQuest, Springer, Scopus, Wiley) for English peer-reviewed journal articles published 2004–2023; initial search yielded 3,506 records, reduced to 2,443 after duplicate removal, then 134 after title/abstract screening, and finally 35 articles.
- Inclusion criteria required K-16 students, STEM subjects, concept map vs. non-concept map comparison, concept maps used as a teaching tool rather than solely for assessment, academic performance as the dependent variable, and sufficient data for effect-size calculation; two independent coders had Cronbach’s alpha = 0.882.
- Effect sizes used Hedges’ g in Comprehensive Meta-Analysis 3.0, with random-effects modeling due to heterogeneity; six moderators were analyzed (subject, educational level, experiment duration, concept map construction mode, learning equipment, pedagogical approach), plus publication-bias tests (fail-safe N = 2,184; Egger p = 0.075; Begg p = 0.680) and sensitivity analyses.
Limitations
- Stated by the paper: only 35 papers were included across two decades; only English-language studies were included; studies lacking sufficient effect-size data were excluded; disciplinary distribution was unbalanced and mainly science; AR, AI-assisted mapping, and generative AI were not analyzed as specific moderators due to too few articles; meta-analytic moderators show associations rather than causal moderation; and the study focuses on achievement while only briefly mentioning motivation, cognitive load, and engagement.
- Evident from the text: the abstract describes “37 studies,” while the methods and results describe 35 articles and 37 extracted effect sizes, creating an inconsistency in how the study count is reported.
- Evident from the text: several moderator categories are very small (e.g., chemistry n = 2, mathematics n = 2, engineering n = 2, technology n = 2, contextual learning n = 2, game-based learning n = 2, > 3 months n = 1), and the paper itself notes limited reliability for some of these conclusions.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
5 Valid · 6 Not valid · 18 Not checked · 76 Own results and interpretations
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The two mathematics studies included in this analysis revealed mixed results.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
In the paper
“Van Orman et al. (2022) found that concept maps were less effective than exemplar-based learning when teaching the algebra skill'Completing the Square' .”
What the sources say
- Not checked — no full text of the sourcecites Van Orman et al. (2022) source
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“In this study, approximately 66% of the experiments (where concept maps were teacher/researcher-generated) lasted less than one week.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“However, the effectiveness of concept maps in physics is not significant in all cases.”
plainSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“Results suggest that interventions lasting between 1 week and 1 month significantly enhance students'achievement”
hedgedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
49 more claims
Create a free account to see all 105 claims and to verify the 1 reference still open.
- It may state two separate claims as one. 30
- Probably not a substantial claim. 28
- The quote only partly states this claim. 1
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.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Concept maps differently influence students'achievement in STEM education across moderator variables Subject · p. 12
“Van Orman et al. (2022) found that concept maps were less effective than exemplar-based learning when teaching the algebra skill'Completing the Square' .”
§ Concept maps differently influence students'achievement in STEM education across moderator variables Subject · p. 12
“Therefore, the effect of the example learning method is inferior to that of the concept map method.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Concept mapping in STEM education · p. 2
“Concept mapping was significantly more helpful to girls than to boys.”
§ Concept maps differently influence students'achievement in STEM education across moderator variables Subject · p. 12
“Bot and Eze (2016) found that the concept mapping method can significantly improve students'achievement, especially that of boys, when teaching trigonometry.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Abstract · p. 1
“This meta-analysis synthesizes findings from 37 studies conducted between 2004 and 2023”
§ Calculating effect sizes · p. 7
“Thus, a total of 37 effect sizes were ultimately extracted from 35 studies.”
Citations
Cites
Baron, R. M., & Kenny, D. A. (1986) · Journal of Personality and Social Psychology · doi:10.1037/0022-3514.51.6.1173 · Cited by 91,532 · cited 1×
“A moderator variable is a discrete or continuous variable that affects the direction and/or strength of the relation between an independent (or predictor) variable and a dependent (or criterion) variable (Baron & Kenny, 1986).”
Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement
Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & The PRISMA Group (2009) · PLoS Medicine · doi:10.1371/journal.pmed.1000097 · Cited by 64,042
Measuring inconsistency in meta-analyses
Higgins, J. P. T., Thompson, S. G., Deeks, J. J., & Altman, D. G. (2003) · BMJ · doi:10.1136/bmj.327.7414.557 · Cited by 63,139 · cited 1×
“It is a method for jointly determining the percentage of inter-study variation that is truly heterogeneous across studies rather than sampling (Higgins et al., 2003).”
Bias in meta-analysis detected by a simple, graphical test
Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997) · BMJ · doi:10.1136/bmj.315.7109.629 · Cited by 56,722 · cited 1×
“Further tests for publication bias, including Egger’s test ( p = 0.075 > 0.05, t = 1.84 < 1.96) and Begg and Mazumdar’s rank correlation test ( p = 0.680 > 0.05, Z = 0.412 < 1.96), confirmed that the likelihood of publication bias is low (Begg & Mazumdar, 1994; Egger et al., 1997).”
InfluentialQuantifying heterogeneity in a meta‐analysis
Higgins, J. P. T., & Thompson, S. G. (2002) · Statistics in Medicine · doi:10.1002/sim.1186 · Cited by 36,763 · cited 1×
“Sensitivity analysis was conducted to detect any potential outliers that might influence the overall effect size (Higgins & Thompson, 2002).”
Operating characteristics of a rank correlation test for publication bias
Begg, C. B., & Mazumdar, M. (1994) · Biometrics · doi:10.2307/2533446 · Cited by 16,955 · cited 1×
“Further tests for publication bias, including Egger’s test ( p = 0.075 > 0.05, t = 1.84 < 1.96) and Begg and Mazumdar’s rank correlation test ( p = 0.680 > 0.05, Z = 0.412 < 1.96), confirmed that the likelihood of publication bias is low (Begg & Mazumdar, 1994; Egger et al., 1997).”
InfluentialBorenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009) · doi:10.1002/9780470743386 · Cited by 14,225 · cited 2×
“Publication bias in meta-analysis arises when studies reporting positive or significant findings are more likely to be published than those with non-significant or negative results (Borenstein et al., 2021).”
Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement
Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2010) · International Journal of Surgery · doi:10.1016/j.ijsu.2010.02.007 · Cited by 12,156 · cited 1×
“The study employed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework (Moher et al., 2010) to guide the search process.”
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2021) · doi:10.1002/9781119558378 · Cited by 9,799
(2000) · Cited by 9,198
Research works printed but not linked: 1
- Haris, A. (2014). Understanding the differences of learners and learning needs through the development of cognitive, language, social, moral and personal. International Journal of Advanced Research.
Cited by
A meta-analysis of AI-enabled personalized STEM education in schools
Li, S., Zeng, C., Liu, H., Jia, J., Liang, M., Cha, Y., Lim, C. P., & Wu, X. (2025) · International Journal of STEM Education · doi:10.1186/s40594-025-00566-y · Cited by 55
“Wang et al. (2025) synthesized 37 studies (2004–2023) in a meta-analysis examining the impact of concept mapping in STEM education and found a medium effect size (g = 0.630).”
InfluentialSystems Thinking in the Role of Fostering Technological and Engineering Literacy
Kurent, B., & Avsec, S. (2025) · Systems · doi:10.3390/systems14010005 · Cited by 15
Dai, Z., Huang, F., Xiong, J., Yang, Y., Zhang, Q., & Peng, X. (2026) · International Journal of STEM Education · doi:10.1186/s40594-026-00631-0 · Cited by 6
“…practice, teachers often promote students’ inquiry-based learning through collaborative inquiry, dialogic teaching, and scaffolded guidance, while also incorporating knowledge-building tools such as concept maps (Lehesvuori et al., 2018; Pedaste et al., 2015; Strat et al., 2024; Wang et al., 2025).”
Полуніна, А., Столяренко, В., Nechypurenko, P. P., & Semerikov, S. O. (2026) · Social Sciences & Humanities Open · doi:10.1016/j.ssaho.2026.102927 · Cited by 1
Xu, Y. (2026) · Progress in Artificial Intelligence · doi:10.1007/s13748-026-00497-x
Metadata
- Authors
- Wang, X.-M., Wang, J.-L., Xu, S.-Y., & Xu, S.-J.
- Affiliations
- Zhejiang Normal University (CN)
- Corresponding author
- Wang, Xiao-Ming
- ORCID
- Wang, Xiao-Ming
- Venue
- International Journal of STEM Education
- Published
- 2025-06-21
- Volume
- 12
- Issue
- 1
- ISSN
- 2196-7822 (ISSN-L)
- OpenAlex
- W4411504293
- Cited by
- 14
- References
- 93
- Citations by year
- 2026: 10 · 2025: 4
- FWCI
- 5.01
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Locations
- International Journal of STEM Education (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- Higher Education Teaching Reform Project of Zhejiang Province (JG-2022-0146)
- Topics
- Social Sciences › Education › Science Education and Pedagogy
- Physical Sciences › Computer Science › Information Systems › Educational Technology and Assessment
- Social Sciences › Psychology › Experimental and Cognitive Psychology › Visual and Cognitive Learning Processes
- Keywords
- concept mapping, STEM education, academic performance, knowledge organization, knowledge integration, science education, technology education, secondary education, inquiry-based learning, contextual learning, teaching methods
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.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Concept mapping in STEM education: a meta-analysis of its impact on students' achievement (2004–2023). https://app.cochat.ai/papers/10.1186/s40594-025-00554-2
- The paper
- Wang, X.-M., Wang, J.-L., Xu, S.-Y., & Xu, S.-J. (2025). Concept mapping in STEM education: a meta-analysis of its impact on students' achievement (2004–2023). International Journal of STEM Education. https://doi.org/10.1186/s40594-025-00554-2
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.