When middle really means “top” or “bottom”: An analysis the 16PF5 using Bock’s nominal response model

Aja Louise Murray, Thomas Booth, Dylan Molenaar

Research output: Contribution to journalArticlepeer-review

Abstract / Description of output

When self-report items with a Likert-type scale include a middle response option (e.g. “Unsure”, “Neither Agree nor Disagree” or “?”), this middle option is assumed to measure a level of the trait intermediate between the high and low response categories. In the current study, we tested this assumption in the 16 Personality Factor Questionnaire, Version 5 (16PF5) by fitting Bock’s nominal response model in the US and UK standardization samples of the 16PF5. We found that in many cases, the middle option was indicative of higher levels of the latent trait than the ostensibly highest response option. In certain other cases, it was indicative of lower levels of the latent trait than the ostensibly lowest response option. This undermines the use of a simple successive integer scoring scheme where responses in adjacent response categories are assigned scores of ‘0’, ‘1’ and ‘2’. Recommendations for alternative scoring schemes are provided. Results also suggested that certain personality traits, especially neurotic traits, are associated with a tendency towards selecting the middle option.
Original languageEnglish
JournalJournal of Personality Assessment
Early online date7 Nov 2015
DOIs
Publication statusPublished - 2015

Fingerprint

Dive into the research topics of 'When middle really means “top” or “bottom”: An analysis the 16PF5 using Bock’s nominal response model'. Together they form a unique fingerprint.

Cite this