Graduation Year

2022

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Curriculum and Instruction

Major Professor

Thomas E. Miller, Ph.D.

Committee Member

Amber D. Dumford, Ph.D.

Committee Member

Jennifer Schneider, Ph.D.

Committee Member

Ruthann Atchley, Ph.D.

Keywords

case management, mental health, predictive analysis, student support

Abstract

Higher education administrators are tasked with supporting and retaining students with increasing needs. These needs often include emotional and mental health issues but can worsen to include suicidality and violence toward others. Traditional campus approaches for supporting students and intervening for violence, such as counseling and campus safety, have been reactionary rather than proactive. Behavioral Intervention Teams (BITs) have emerged as a mechanism for heading off violence before it occurs while also supporting students who may never engage in violence but need support. These teams were born out of the concept that violence is preventable and have grown into a strategy for student support and retention.

Although BITs have existed for over a decade, that they remain reactive, rather than proactive. BITs rely on community members to make a referral to the team for a student who is experiencing difficulty. This is a reactive approach, as it requires a student to already be in distress before receiving a referral for assistance. To truly be preventative, BIT administrators need the ability to predict who is likely to need a support before the individual is in distress.

To assist in developing a predictive model for BIT referrals, this study aimed to explore expected student engagement as a potential predictor that a student will need assistance. This study was conducted at a large public research institution and included first time in college (FTIC) students who enrolled between 2015 and 2019. The study used the Beginning College Survey of Student Engagement to explore the relationship between student engagement and a referral to the BIT. The BCSSE scales therefore served as the predictor, or independent, variables.

Specifically, the BCSSE scales used as the predictor variables in this study included Expected Engagement in Collaborative Learning, Expected Discussions with Diverse Others, Expected Academic Perseverance, Expected Academic Difficulty, and Importance of Campus Environment. The outcome, or dependent variable for this study, was measured by whether a student was referred to the institution’s BIT at any point during the time frame for this study. Additionally, gender and race were included as covariates. Given that the predictor variables in this study were continuous and the outcome variable dichotomous, logistic regression was used to conduct the analysis.

The logistic regression demonstrated no significant relationships between individual BCSSE scales and a referral to the BIT. There was a significant relationship between expected academic perseverance and a referral to the BIT (p = .010) when the five BCSSE scales were included as a group of predictors. Finally, the relationship between expected academic perseverance and a referral to the BIT remained statistically significant (p = .020) when included in a model with the five BCSSE scales and the covariates. Additionally, students who identified their gender as another gender identity and/or preferred not to respond regarding gender were more likely to be referred to the BIT than men, and students who identified their race as Black or African American, multiple races, or preferred not to respond regarding race were significantly more likely to be referred than White students.

Despite statistically significant relationships between these variables and a referral to the BIT, this study did not lead to a predictive model for a referral to the BIT. The logistic regression models all predicted that 100% of students would not need a referral and therefore could not accurately predict a referral to the BIT.

Download

Share

COinS