Attitudes toward Using and Teaching Confidence Intervals: A Latent Profile Analysis on Elementary Statistics Instructors

Abstract

The use of confidence intervals (CIs) for making a statistical inference is gaining popularity in research communities. To evaluate college statistics instructors’ readiness to teach CIs, this study explores their attitudes toward teaching CIs in elementary statistics courses, and toward using CIs in inferential statistics. Data were collected with a survey that classifies instructors’ attitudes on the basis of three previously established pedagogical components: affective, cognitive, and behavioral. Based on the survey responses from 270 participants, we created three profiles (subgroups) via latent profile analysis, and identified each profile’s pattern of attitudes toward CIs and common characteristics of the instructors that fit each profile. In addition, we compared the profiles across groupings created by six variables: gender, academic background, statistics teaching experience, subject preference, degree level, and desire to improve teaching. The results of the latent profile analysis support three profiles within the population of statistics instructors, and the results of the comparative analysis of teacher characteristics indicate that the six variables are moderate to strong predictors of the grouping of the sample into three profiles.