Portfolio assessment, which can be defined as a purposeful collection of a student’s work that tells a story of the student’s efforts, progress, or achievement in one or more areas (Arter & Spandel, 1992), has received scholarly attention. Depending on various purposes, Miller et al. (2009) summarize portfolio use in four dimensions as (a) instruction and assessment, (b) current accomplishments and progress, (c) showcase and documentation, and (d) finished and working (p. 293), which do not only show the instructional and evaluative value of portfolio use, but also highlight the formative contributions to self-reflection, higher-order thinking, and self-regulation. Either paper-based or electronic, the use of portfolios has been shown to improve several variables in related research. The redundancy of findings specifically related to the effects of portfolio assessment on student achievements and attitudes, as the most frequently examined dependent variables, suggests a systematic review not only to interpret the magnitude of influence in terms of effect size, but also to facilitate scholars in designing their future studies. The following research questions, with achievement and attitude as the dependent variables and portfolio, either paper-based or electronic, as the independent variable, were addressed in this study:

(1)   What is the influence of portfolio use on student achievement?

(2)   What is the influence of portfolio use on student attitude?

(3)   How do several moderator variables (year of publication, type of publication, education level, subject, and type of portfolio) influence the effects of portfolio use?

One novel contribution of the present meta-analysis is that it provides a review of extant research examining the effect of portfolio use on achievement and attitude in the context of Türkiye by including more studies proposed by search of related database and it quantifies the findings related to several moderators.

This research used meta-analysis to address questions pertaining to effects of portfolio assessment on student achievements and attitudes in scholarly work in Türkiye. For research purposes, experimental or a quasi-experimental research examining the effects of portfolios assessment on student achievement and attitude were included in analyses. The rigorous search of the premium databases in Turkiye yielded a total of 62 studies with 62 effect sizes on achievement (N=3.664) and 30 effect sizes on attitudes (N=1.779). Hedges’ g, a less biased measure of effect size (Lambert & Alhassoon, 2015), was used to determine the extent of the difference between achievement and attitude scores of experimental and control groups. As suggested by Field and Gillett (2010) for studies of meta-analysis in social sciences, random-effects model was employed for the analyses. Computations were carried out using Comprehensive Meta-Analysis Version 4 (Borenstein et al., 2022). The robust review of related research, the use of inclusion and exclusion criteria, and the reporting of effect sizes ensured the validity of the study. The reliability was provided with the agreement between two coders with a doctoral degree in education, who carried out the selection of studies, analysis and interpretation of data. Several statistics were employed to test the data for heterogeneity and publication bias (Borenstein et al., 2021). I2 and Q statistics were performed in the measurement of heterogeneity. The significance level of Q value (e.g. p<.01) and I2 values of 25%, 50% and 75% that represent low, moderate and high heterogeneity, respectively (Higgins & Thompson, 2002) were used to detect heterogeneity. Publication bias was examined with Egger’s Regression Test, Rosenthal’s FSN, and Duval & Tweedie’s Trim and Fill (TFM). The symmetry observed in the funnel plots and the nonsignificant p values (>.05) in Egger’s Regression Test would show the nonexistence of publication bias. For the Rosenthal’s FSN, the 5k+10 value (Fragkos et al., 2014; Mullen et al., 2001) was used as a cut-off point to observe the nonexistence of publication bias. Analog to ANOVA was used to explore the moderating effect of year of publication, type of publication, education level, subject, and type of portfolio on mean effect sizes in studies examining the effect of portfolio assessment on both achievement and attitude.

The random effects model reflected the mean effect size as 0,998 (SE = 0,11, 95% CI = [0,79, 1,21]), which implied a large positive effect of portfolio assessment on students’ achievements; whereas the mean effect size of 0,676 (SE = 0,10, 95% CI = [0,48, 0,88]) for studies examining the influence of portfolio assessment on student attitudes implied a medium positive effect (Cohen, 1988). Further analyses for both achievement and attitude showed that data were highly heterogeneous and displayed no clear evidence of publication bias. A cumulative analysis of all the Analog to ANOVA results regarding the moderators in both achievement and attitude studies revealed the significant effect of subject only. In other words, subject as a variable was found to be the source of variance moderating the magnitude of effect sizes significantly in both achievement and attitude studies.

Arter, J., & Spandel, V. “Using Portfolios of Student Work in Instruction and Assessment.” Educational Measurement: Issues and Practice, 1992, 11, 36–44. Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2021). Introduction to Meta-Analysis (Second ed.). Wiley. Borenstein, M., Hedges, L. E., Higgins, J. P. T., & Rothstein, H. R. (2022). Comprehensive Meta-Analysis Version 4. In Biostat, Inc. www.Meta-Analysis.com Field, A. P., & Gillett, R. (2010). How to do a meta-analysis. British Journal of Mathematical and Statistical Psychology, 63(3), 665-694. https://doi.org/10.1348/000711010X502733 Fragkos, K. C., Tsagris, M., & Frangos, C. C. (2014). Publication bias in meta-analysis: confidence intervals for Rosenthal’s fail-safe number. International scholarly research notices, 2014. https://doi.org/10.1155/2014/825383 Higgins, J., & Thompson, S. G. (2002). Quantifying heterogeneity in a meta-analysis. Statistics in Medicine, 21(11), 1539-1558. Lambert, J.E., & Alhassoon, O.M. (2015) Trauma-focused therapy for refugees: Meta-analytic findings. Journal of Counseling Psychology, 62(1), 28–37. https://doi.org/10.1037/cou0000048 Miller, M.D., Linn, R.L., & Gronlund, N.E. (2009). Measurement and assessment in teaching (10th ed). Pearson Education. Mullen, B., Muellerleile, P., & Bryant, B. (2001). Cumulative meta-analysis: a consideration of indicators of sufficiency and stability. Personality and Social Psychology Bulletin, 27(11), 1450.