Date of Award


Document Type


Degree Name



Quantitative Research Methods

First Advisor

Elizabeth Anderson

Second Advisor

Kathy Green

Third Advisor

Duan Zhang

Fourth Advisor

Judy Kiyama


Measure, Multi-faceted Rasch modeling, MFRM, Retention, Social support


Being born into circumstances of low-income, having a racial minority status, and/or non-college educated families dwindle the opportunities for many students to obtain a college degree (Cox, 2016; Engle & Tinto, 2008; Jenkins et al., 2013). While many institutions of higher education have diligently worked to develop programs geared towards attending the educational inequalities among diverse student populations, there is still a great need for programs centered on the inequalities surrounding social support (Cox, 2016; Ward et al., 2012; Soria & Stebleton, 2012).

The purpose of this study was to develop and assess a measure to examine perceived social support for undergraduate students using the Culturally Engaging Campus Environments (CECE) model of college success as a framework for developing social support items that reflected specific events related to the college experience and include questions specific to friend/classmate, family, and college faculty.

Using confirmatory factor analysis and multi-faceted Rasch modeling, the researcher found that the measure demonstrated reliable model fit and was invariant across a grouping variable designed to examine the relationship of the number of marginalized identities and social support. The 3-factor model fit indices indicated that the model had an excellent fit (χ2 (374) = 833.77, p< .001, RMSEA = 0.069, CFI = 0.911, AIC = 955.77) and the factor loadings were all above the .70 cutoff, whereas the 1-factor model indices did not meet either of the suggested fit indices thresholds (χ2 (378) = 3057.51, p < .001, RMSEA= 0.165, CFI = 0.472, AIC = 3171.51). The results were further supported by running a χ2 difference test (χ2 (4) = 2223.74, p < .001). MFRM results indicated that there was not a significant difference between the groups in logit position for the family factor (χ2 (2) = 4.8, p = .09, SD = .07, Separation = 1.23, Strata = 1.97, Reliability = .60) or the friend/classmate factor (χ2 (2) = 4.6, p = .10, SD = .04, Separation = .64, Strata = 1.19, Reliability = .29). However, there was a significant difference between the groups for the faculty factor (χ2 (2) = 7.1, p = .03, SD = .07, Separation = 1.24, Strata = 1.98, Reliability = .60).

Publication Statement

Copyright is held by the author. User is responsible for all copyright compliance.


Received from ProQuest

Rights holder

Heather M. Blizzard

File size

231 p.

File format





Education, Social research