Examining the Factor Structure of the BIMAS-2 SEL Scales

In February 2025, Sequoyah Bell and Allison Dembowski (project coordinator) led a poster presentation at the annual convention of the National Association of School Psychologists in Seattle, WA. Co-authors included Christy Walcott, Brandon K. Schultz, and Mark D. Weist (University of South Carolina).

Introduction

BIMAS-2

The BIMAS-2 (McDougal et al., 2011) is a robust universal behavior screener and progress monitoring tool. Five norm-referenced scales are derived on the Standard Scale: Conduct, Negative Affect, Cognitive/Attention, Social, and Academic Functioning.

Creation of SEL Scale

The BIMAS-2 creators mapped items from the standard scale onto the five domains of social-emotional learning (SEL) in the CASEL framework (CASEL, 2024) (see graphic) to form the BIMAS-2 SEL Scales: Self-Awareness, Self-Management, Responsible Decision-Making, Relationship Skills, and Social Awareness (EduMetrisis, 2021).

Present Study

The original SEL scales were never empirically tested. We examined internal reliability of the original BIMAS-2 SEL Scales. Additionally, many Items load onto multiple SEL scales rendering intercorrelated scales. An exploratory factor analysis (EFA) of all 33 BIMAS-2 SEL items could reveal the factor structure. Research questions were:

•Are the original BIMAS-2 SEL scales reliable?

•Does EFA discover factors representing all 5 CASEL SEL areas consistent with the original BIMAS-2 SEL scales?

•Do the 33 items fit well on single factors?

Methods

•Teachers completed the BIMAS-2 for all students in their class, over 3 consecutive semesters (4th & 5th grade) at 16 schools in NC and SC as part of a larger ongoing study.

•Spring ‘23 dataset was used (n = 923).

•Reliability was based on the full sample. Then, we randomly split the sample in half, and one half was used for the EFA (n = 463).

Results

Exploratory and Confirmatory Factor Analysis: BIMAS SEL items Spring ‘23

Extraction method was principal axis factoring with oblique rotation (Promax with Kaiser Normalization). Factor loadings above .30 presented. Two items from BIMAS-2 SEL scales did not adequately load onto any factor (Spoke Clearly to Others & Shared What He/She was Thinking About). Results presented in the Table below.

28, TLI = .91653,0.063], SRMR =X2(182)=463, p =0.000, RMSEA = .058,90% [0.053,0.063], SRMR =0.052, CFI =.928, TLI = .9160.052, CFI =.928, TLI = .916

NOTE: Items load onto original BIMAS SEL scale + new EFA factor. (original BIMAS SEL scale on which item loads.) *Wording of this factor matches CASEL-5 framework.

Discussion

Exploratory Factor Analysis

Our EFA identified 5 distinct factors; 4 mapped well conceptually onto the CASEL-5 framework, but one area (Social Awareness) was not represented. Many of the original BIMAS SEL items for Soc.A. fell on the new EFA Relationship Skills factor and a few others on Responsible Decision-Making.

Confirmatory Factor Analysis

Our CFA confirmed the final 4 distinct factors that mapped well conceptually onto the CASEL-5 framework. According to the model factors that should be observed to indicate whether the model is a good fit, the 4 factors were a good fit for the model. The final included factors were Responsible Decision Making, Self-Management, Self-Awareness, and Relationship Skills. These factor analysis tests the validity of the BIMAS-2 measures and confirms that 4 out of 5 of the original BIMAS-2 SEL framework categories have strong validity.

Once specific model indices were observed, one item “Had Problems Paying Attention” was considered to have a high covariance with another item. Once those modifications were made, the model fit was significantly better than the previous original model and our final factors and factor loadings were confirmed.

Future studies involving the empirical analysis of the BIMAS-2 should include testing for invariance across time and across racial & gender groups for the other time set data samples that were collected from the ESSS project using the BIMAS-2 data.

Additionally, a CFA should be conducted on the original loaded factors of the BIMAS-2 (suggested by the authors) and compared to our BIMAS-2 model to evaluate statistically which model fit is better to use for future uses of the BIMAS-2 SEL scale. 

Acknowledgements

The research reported here was supported by the Institute of Education Sciences, U.S. Department of Education, through Grant R324A210179 to East Carolina University. The opinions expressed are those of the authors and do not represent views of the Institute or the U.S. Department of Education. Special thanks to Dr. Alex Schoemann at ECU.