Psychometric properties of the college survey for students with brain injury: Individuals with and without traumatic brain injury

Abstract

Purpose: The psychometric properties of the college challenges sub-set from The College Survey for Students with Brain Injury (CSS-BI) were investigated with adults with and without traumatic brain injury (TBI).

Methods: Adults with and without TBI completed the CSS-BI. A sub-set of participants with TBI were interviewed, intentional and convergent validity were investigated, and the internal structure of the college challenges was analysed with exploratory factor analysis/principle component analysis.

Results: Respondents with TBI understood the items describing college challenges with evidence of intentional validity. More individuals with TBI than controls endorsed eight of the 13 college challenges. Those who reported more health issues endorsed more college challenges, demonstrating preliminary convergent validity. Cronbach’s alphas of >0.85 demonstrated acceptable internal reliability. Factor analysis revealed a four-factor model for those with TBI: studying and learning (Factor 1), time management and organization (Factor 2), social (Factor 3) and nervousness/anxiety (Factor 4). This model explained 72% and 69% of the variance for those with and without TBI, respectively.

Conclusion: The college challenges sub-set from the CSS-BI identifies challenges that individuals with TBI face when going to college. Some challenges were related to two factors in the model, demonstrating the inter-connections of these experiences.

• Academics
• cognition
• college
• post-secondary

Acknowledgements

The authors wish to acknowledge Jocelyn Yu, MA, for her contributions to data coding and to the many individuals with TBI who shared their college challenges with us.

Notes

Notes

1. Note that the CSS-BI also includes questions about rehabilitation, on-campus services and life changes, but responses to these are not included in this analysis.

2. Planned comparisons of health issues and college challenges were performed between the demographic and non-demographic sub-groups. There were no significant differences (p ≤ 0.05) or trends (p ≤ 0.10). Therefore, data were collapsed across sub-groups, for all further analyses.