Add your reactions, reflections, and questions to the discussion questions here! by Claire Chuter

SEL Interventions

2025-01-08 19:53:14 UTC

What are the Programs that are included in these 12 data sets and if there are programs that are represented more than once are they weighted accordingly?

Supplementary Tables

2025-01-08 19:54:47 UTC

I saw the note about the mod table and also noted you had screenshot the under review draft table from the under review paper. Here is the mod table

Here is the final published paper with tables

2025-01-08 19:55:54 UTC

ks and ns

2025-01-08 19:57:00 UTC

what are the ks and ns for each group? Is there a place I can review those before Friday? Thank you!

Effect size estimates

2025-01-09 12:43:24 UTC

Would love to see these caterpillar plots with the effect size labeled.

Moderators

2025-01-09 12:46:37 UTC

Think that the most compelling moderators for the SEL field are the minutes of student facing instruction and number of discrete sessions. The most compelling moderator for the meta-analytic methods field is control group type.

Comparison to Cipriano

2025-01-09 12:51:28 UTC

Perhaps a more appropriate comparison to Cipriano would be just the effects for the US? That would require re analyzing Cipriano....

Negative Effects

2025-01-09 12:58:50 UTC

For some groups on some outcomes, SEL programs may have negative effects

Width of the prediction intervals

spekospicer — 2025-01-10 17:25:43 UTC

I'm curious, why are the prediction intervals for some of these groups so small (e.g., mixed race students)?

interesting findings and questions

2025-01-10 17:32:03 UTC

interesting: 1) no effect/negative effects for Hispanic and African American youth outcomes on family relationships and prosocial behaviors. 2) positive effect for discipline and school functioning for Hispanic, African American, and Native students. Curious how SEL may have impacted school staff on decisions that influence discipline and school functioning. 3) negative effect for family relationships and prosocial behaviors for White students;

Questions: 1) why does graph for Native Hawaiian students not have 95% CI? 2) were discipline outcomes not available for different SES variables? 3) what influenced the CI spread variability?

SEL Skills

spekospicer — 2025-01-10 17:35:08 UTC

It's interesting that for each group, the effects on SEL skills is quite small or negative. I.e., it's at the bottom of the pack in each caterpillar plot.

2025-01-10 17:39:45 UTC

Are all of the outcomes in one model? Or are these separate models for each outcome?

Hard to resist comparing specific groups to one another

2025-01-10 17:40:58 UTC

Especially when the plots are next to one another. Something to consider for the manuscript

Plot's vertical column

cheyeonha — 2025-01-10 17:44:35 UTC

The outcomes in the vertical columns are sorted differently across subgroups, making it difficult to compare the results at a glance. It might be helpful to visualize the results in the same domain order. OR can we show the results by combining one plot (showing the subgroup results together in each domain)

Moderators

spekospicer — 2025-01-10 17:46:19 UTC

Claire: Based on the conversation about SEL skills and measurement --> maybe we should determine what measures are used and see if that moderates effects?

Showing the sample size of each subgroup (student N)

cheyeonha — 2025-01-10 17:47:25 UTC

The number of samples in each group is expected to differ greatly. This may be influencing the intervals displayed in the graph. Including the student sample sizes alongside the table would be beneficial.

What is the appropriate data set to compare to, to offer context?

spekospicer — 2025-01-10 17:48:07 UTC

Josh: Could we compare each the effects observed here to the effects observed in the U.S.-based analysis from the Yale team study?

Comparing to the US dataset

2025-01-10 17:55:11 UTC

Would be interesting to see how these IPD estimates compare to marginal means for race/SES categories from the Yale team’s dataset if those are available

A need for descriptive analysis of the studies

spekospicer — 2025-01-10 17:55:33 UTC

What are the programs? What theories ground those programs? What measures were included in each domain?

Theory about race and/or SES and the different outcome domains

spekospicer — 2025-01-10 17:57:08 UTC

What does theory suggest our hypotheses should be about these relationships, within group? How does that relate to the measurements used?

How to think about this sample of studies?

spekospicer — 2025-01-10 18:02:42 UTC

Rebecca: We should be careful about positioning these effects as the "truth" in relation to what we observed from a broader group of studies. Rather, we should be thinking of this sample as a way to dig into the narrative -- what are the possible explanations that are coming up from these data?

Great job Claire and team!

2025-01-10 18:03:54 UTC

This is really fascinating work!

SEL Skills

2025-01-10 18:04:59 UTC

In my dissertation I look at the actual measures included in the SEL Skills domain and find a lot of heterogeneity. It would be great to look at the outcomes that make up SEL Skills in your data. Happy to help if that would be useful!