Psychometric properties of the Statistical Anxiety and Attitudes Towards Statistics subscales of the HFS-R* revisited

Tine Nielsen, Svend Kreiner

Research output: Contribution to conference without a publisher/journalAbstractResearchpeer-review

Abstract

The HFS-R* (NK-2018) was developed to obtain a contemporary and valid Danish questionnaire for assessing attitudes and relationship towards statistics for use with higher education students taking statistics courses within other academic disciplines. The HFS-R consists of four subscales totalling 26 items; Test & Class Anxiety (TCA), Interpretation Anxiety (IA), Fear of Asking for Help (FAH) and Wort of Statistics (WS). In NK-2018, item analyses were conducted using Rasch models with a sample of Sociology and Public Health students in their first or second statistics course in university. The aim of this study was to conduct a second investigation of construct validity and psychometric properties of the HFS-R subscales with a student sample from another academic discipline to see if previous results could be replicated. We used the same Rasch measurement models to investigate whether the psychometric properties (including fit of individual items, overall homogeneity, differential item functioning, reliability and targeting) of the HFS-R in this study were comparable to the previous results. Our study sample consisted of two consecutive year-cohorts of first semester Psychology students in their first statistics course (N = 341) – thus we could not conduct DIF-analyses for first/second statistics course. Data were collected in class using a paper-pencil questionnaire one month into the courses. The current results only differed slightly from the previous results: we found evidence of age DIF in the TCA subscale of the same magnitude and identical evidence against local independence of two item pairs as NK-2018. For the IA subscale, we found no evidence against fit to the Rasch model, where NK-2018 found two items lacked local independence. For the FAH subscale, we found evidence against local independence of four items, where NK-2018 found no evidence against fit to the Rasch model. The biggest discrepancy between studies, was evidence against the fit of one item in the WS subscale as well as evidence against local independence between two WS items, whereas NK-2018 found no evidence of either. We have thus strengthened the evidence that the HFS-R subscales can be considered essentially valid and objective. * HFS-R = “Holdninger og Forhold til Statistik – Revideret” in Danish
Original languageEnglish
Publication date3 Nov 2021
Publication statusPublished - 3 Nov 2021
EventNordic Educational Research Association (NERA) - 2021 - Odense, Odense, Denmark
Duration: 3 Nov 20215 Nov 2021

Conference

ConferenceNordic Educational Research Association (NERA) - 2021
LocationOdense
Country/TerritoryDenmark
CityOdense
Period03/11/2105/11/21

Cite this