Comparison of self-administered versus read-aloud questionnaires for psychological measurement in students with low intellectual functioning: use of frequentist and Bayesian approaches

Abstract

Background Students with low intellectual functioning (LIF) often experience barriers to participating in social research due to the literacy demands of the survey’s typical self-administered format. Although evidence for the validity of the read-aloud format for educational testing abounds, few studies have analyzed the impact of application formats on attitudes or opinion questionnaires for LIF students. Aim To analyze the effect of self-administered vs read-aloud formats on LIF and typical development (TD) students using four psychological questionnaires for school contexts (Student Engagement Instrument, Multidimensional School Engagement Scale [MSES], Brief Multidimensional Student Life Satisfaction Scale, and School Participation Scale). Method A mixed factorial (2x2) design was used. Thirty-two students participated (14 to 19 years old; M = 15.39; SD = 1.27): 17 with LIF and 15 with TD. Results Reliability indices between formats for LIF students in most questionnaire subscales were found to be adequate and equivalent. All instrument subscales had appropriate intra-subject correlations between formats, indicating that LIF students had similar scores in both. Only the MSES showed a format effect, where LIF students reported fewer disengagement behaviors in the read-aloud format. Frequentist and Bayesian statistics were conducted looking for convergences due to the small sample size. Conclusion We discuss the case-related appropriateness of each application format and propose a new criterion to choose between them to guarantee the inclusion of LIF students in psychological research.