Investigation of Bivariate Grid-Type Items for Measuring Attitudes in Statistics Education: Preliminary Results
Abstract
Likert-type items are ubiquitous in attitude research in statistics education but imply a reciprocal relationship between positivity and negativity in the construct being measured. Based on historical challenges measuring some constructs in a widely used framework in statistics education (Eccles’s Expectancy-Value Theory [EVT]), we speculate that the reciprocal relationship implied by the Likert-type items may not be appropriate. Evaluative Space Grid (ESG) items have been proposed as an alternative: respondents indicate their positivity and negativity on a grid that does not impose a reciprocal relationship. However, there have been relatively few studies that focus on ESG items. This presentation reports on a set of preliminary studies that seek to describe the psychometric properties of ESG items and document evidence of their appropriateness (or lack thereof) for measuring EVT constructs. Data have been collected from introductory statistics students and a general participant pool.
