卓越實證概述 Best Evidence in Brief
Effects of friendships on academic performance and emotion

Friendships play an important role in shaping children’s developmental outcomes. For instance, research has shown that children who have friends and those have no friend differ in various aspects, such as peer group status and academic performance. Chen, along with collaborators, conducted a one-year longitudinal study using an action-partner interdependence model (APIM) to explore the role of friendship in the development of academic achievement and depression for same-domain as well as cross-domain effects, i.e., the associations between academic performance and depression in one member of a friendship dyad and later the academic performance and level of depression of the partner member.

Academic achievement and depression are considered as two domains, thus, an association between antecedent academic results and later academic achievement indicates same-domain effect, while cross-domain refers to an association between academic achievement and depression. The effect within the same individual from Time 1 (T1) to Time 2 (T2) is described as actor effect, and the effect between dyads (two children) is partner effect.

The sample used was 1122 grade 4 and grade 5 students (494 boys) with 561 same sex dyads identified as stable friendships from 61 classes in 5 public elementary schools in Southeast China based on children’s best friends nomination in a two-wave data collection. Students’ academic achievement was an index formed by mean exam scores of Chinese, mathematics, and English adjusted with teacher-rated scores of learning problem. Students’ depression was assessed using the self-reported Childhood Depression Inventory. After controlling for gender, grade, and interdependence between friend dyad members (i.e., all possible same-domain and cross-domain effects of both actor and partners are included in the same analysis model), the results were as follows:

  • A cross-domain actor negative effect of depression on academic performance was found (β =-0.10), i.e., the higher T1 depression the lower T2 academic achievement within same child. However, the reverse effect was not found.
  • A same-domain partner positive effect was found for academic performance (β = +0.14) but not for depression (β =-0.01).
  • A cross-domain partner negative effect was found for T1 academic performance on T2 depression (β =-0.10). In other words, students with an academically better friend had a low level of depression a year later. But a partner effect of T1 depression on T2 academic performance was not found.

The findings suggested that children who had academically competent friends were likely to benefit from the relationship during the development stage for strengthening their academic performance and for reducing their feelings of depression.

 

Source: Chen, X., Zhou, J., Liu, J., Li, D., & Liu, S. (2022). Academic performance and depression in Chinese children: Same-domain and cross-domain effects in friendships. Child Development, 00, 1–15. https://doi.org/10.1111/cdev.13864Read the rest

Language skills may help children to make more friends

Early peer experiences have impacts on children’s academic and social developmental trajectories, and successful peer experiences are important in kindergarten because it sets the foundation for future peer interactions and friendships. Though numerous factors can affect peer experiences, language skills and expressions may influence a child’s ability to initiate, engage, and sustain social activities with their peers.  As it is likely that language skills influence children’s classroom-based friendships, a recent study by Chow and his colleagues explored the difference between children at risk for special language impairment (SLI) and their typical peers regarding friendship centrality and reciprocity in kindergarten classrooms using social network analysis.

Data from a sample 419 children (53% boys, average age = 6.8 years) from 21 kindergarten classrooms in four elementary schools in the Mid-Atlantic region of the United States were collected. Children at risk for SLI (n=190) were identified using scores of the screening tools of the Clinical Evaluation of Language Fundamentals (CELF-5). Peer nomination technique was employed to locate the friendship network in that each child was asked to name three best friends in their classroom. The friendship network centrality was a measure of children’s social position in the classroom, and it was an index generated from number of friends nominated and number of received nominations taking into account classroom size. High centrality scores suggest that a student is more central to the social network. Reciprocity was a dichotomous variable: reciprocity = 1 if a student received at least one reciprocal nomination; reciprocity = 0 if no reciprocal friendship.

The results of Multilevel modelling indicated that:

  • Of the 419 children, 15.2% did not make any best friend nomination, and 16.9% received no nomination. Of the 190 children at risk, 43.0% had reciprocal friendships while for children not at risk, 57.5% had reciprocal friendships.
  • CELF-5 language scores were associated with centrality, and students at risk for SLI had lower centrality than children not at risk.
  • Among students at risk for SLI, boys had significantly lower centrality compared to girls.
  • Though SLI scores were significantly associated with reciprocity, students at risk for SLI were found to have 49% lower chance of having a reciprocal friendship compared to children not at risk.

Although some important variables were not included in the analysis model at both classroom context and individual levels (e.g., verbal intelligence, social skills, English language proficiency), the current study still provides some preliminary evidence of the association between language skills and peer friendship, and girls were significantly more central than boys among children at risk for SLI. 

 

Source: Chow, J. C., Broda, M. D., Granger, K. L., Deering, B. T., & Dunn, K. T. (2021). Language skills and friendships in kindergarten classrooms: A social network analysis. School Psychology, 37(6), 488–500.  https://doi.org/10.1037/spq0000451Read the rest

Linking universal pre-K to college enrollment

Although many studies have shown short term benefits to early childhood education, the long-term benefits have been less explored. A new study by William Gormley Jr. and colleagues found that children who had participated in the Tulsa universal Pre-K program in 2006 were 12% more likely to enroll in a 2-year or 4-year college. In particular, Black and Hispanic children who had attended the Pre-K program were more likely to enroll in 4-year institutions. Children who had participated in the Tulsa Head Start program were 7% more likely to enroll in any college or university, but these results were only slightly statistically significant.

The research team calculated propensity scores to compare the groups of children who did and did not attend the Pre-K program to account for other factors that may have influenced likelihood to enroll in college.

Tulsa’s universal Pre-K program is one of the oldest in the US. Teachers in the program have college degrees, are certified in early childhood education, and are paid similarly to public school teachers. Tulsa also has an effective system of magnet schools that students can attend on the journey from Pre-K to college. These factors may have made the impact of universal Pre-K on Tulsa’s children particularly long-lasting. However, since many studies have shown the increased earning power of having a 4-year college degree, 2-year degree, or even having completed some college credits without completing a degree program, the authors suggest that a strong universal pre-K program could possibly be a worthwhile investment.

 

Source (Open Access): Gormley, W. T., Amadon, S., Magnuson, K., Claessens, A., & Hummel-Price, D. (2023). Universal pre-K and college enrollment: Is there a link? AERA Open, 9, 23328584221147892. https://doi.org/10.1177/23328584221147893Read the rest

Study of a standards-based grading system

A study posted on preprint archive by Kramer and colleagues examined whether 9th grade students’ academic performance and motivation in mathematics improved by using a formative grading system called PARLO (Proficiency-based Assessment and Re-assessment of Learning Outcomes), and compared it to the traditional summative-grading system.

The PARLO system aims to increase academic engagement and achievement by centering instruction, formative feedback, and grading around 10-15 Learning Outcomes per semester and employing a grading system that encourages reassessment and full credit for mastery of content at each learning outcome. A student’s final grade is calculated from the number of “proficient” or “high performance” scores taken from their best assessment at each learning outcome. This differs from traditional grading, which is often based on one-time assessments and summary judgments of student performance.

The mixed-methods, cluster randomized study assigned a diverse set of schools to implement PARLO for two years in their 9th grade algebra and geometry classes or to the control condition, where algebra and geometry teachers attended introductory PD sessions covering formative assessment instruction with no obligation to incorporate new strategies. The analytic sample consisted of 2,736 students (1649E, 1087C), in 65 classrooms (38E, 27C) from 19 schools (14E, 15C).

Quantitative measures (including adapted versions of the Virginia Standards of Learning Multiple Choice Algebra and Geometry Tests) and qualitative interview/survey measures showed that the PARLO system had a positive impact on 9th graders’ learning of mathematics, with an estimated effect size of +0.33 SD on the end-of-course algebra and geometry tests. The program showed similar results across race, gender, and prior achievement. Qualitative data suggest that the PARLO system increased students’ engagement, provided effective strategies to strengthen understanding of mathematics content (e.g., by asking better questions), and encouraged peer-to-peer support. The formative feedback coupled with opportunities to retake assessments led to more chances to review topics over time, which supported deeper connections across mathematical topics. Finally, results suggested that the PARLO system is especially effective for students already engaged and motivated to learn.

Note: A preprint (or e-print) is a completed manuscript that is made publicly accessible prior to formal journal peer review.

 

Source (Open Access): Kramer, S., Posner, M. A., Lawrence, N., Browman, A. S., Krier, K., & Roemer, J. (2021). The impacts of a standards-based grading system emphasizing formative assessment, feedback, and re-assessment: A mixed methods, cluster randomized control trial in ninth grade mathematics classrooms. EdArXiv (Preprint archive). https://doi.org/10.35542/osf.io/pzc3fRead the rest

The first meta-analysis of global learning loss

Almost three years since the COVID-19 pandemic, there is a growing number of studies examining the learning loss of school-aged children during the pandemic. The first meta-analysis of global learning loss has been published in Nature Human Behavior, and reviews the existing evidence on the extent to which learning progress has slowed down since the onset of the pandemic, and how it varies across different groups of students, subject domains and country contexts.

To be eligible for inclusion, studies had to: (1) measure learning progress using test scores that could be standardized across studies using Cohen’s d, (2) base their estimates on empirical data collected since the onset of the COVID-19 pandemic, rather than making projections based on pre-COVID-19 data, and (3) be written in English, Danish, Dutch, French, German, Norwegian, Spanish or Swedish.

A total of 42 studies (including 291 estimates) across 15 countries met these inclusion criteria, conducted from the start of the pandemic in 2020 through March 2022. This meta-analysis revealed that: (1) Overall, a substantial learning loss (Cohen’d = -0.14) was found, meaning that students lost out on 35% of a school years’ worth of learning; (2) On average, learning loss arose early in the pandemic and persisted over time; (3) Learning loss was particularly large among students from low socio-economic backgrounds; (4) Learning loss was larger in math than in reading, and (5) Learning loss was larger in poorer countries.

Authors made suggestions for future research, including how the pandemic affected gender inequality and examining effectiveness of interventions for recovering learning loss. They further recommended that policy initiatives to compensate learning loss need to: (1) prioritize support for students from low socio-economic backgrounds, (2) prioritize support in math and other science subjects, and (3) identify and realize opportunities to complement and expand on regular school-based learning.

 

Source (Open Access): Betthäuser, B. A., Bach-Mortensen, A. M., & Engzell, P. (2023). A systematic review and meta-analysis of the evidence on learning during the COVID-19 pandemic. Nature Human Behaviour, 1–11. https://doi.org/10.1038/s41562-022-01506-4Read the rest