Off-campus UNL users: To download campus access dissertations, please use the following link to log into our proxy server with your NU ID and password. When you are done browsing please remember to return to this page and log out.
Non-UNL users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
A comparison of the psychometric properties of negatively and positively worded questionnaire items
Abstract
The reliability, factor structure, and validity of responses from positively and negatively worded items were compared across two data collection conditions: structured and unstructured. The negatively worded items, which were formed by adding the word “not” to each sentence, were found to be less reliable than the positively worded items across both data collection conditions. Contrary to previous findings, the negatively worded items did not influence the factor structure or validity of the responses. However, it was found that a model including only a trait factor fit the responses better when the data collection was structured than when it was unstructured. The lack of a strong negatively worded item method factor precluded the examination of the relationship between this type of factor and cognitive ability. An examination of the effect of the presence of negatively worded items on the responses to positively worded items found that the type of data collection influenced the existence of this effect. Specifically, positive items presented with negative items were less reliable than positive items presented with other positive items if the data collection was unstructured. When the data collection was structured this effect disappeared. The decrease in reliability has implications on recommendations to include negative items on instruments to guard against response sets but not include them in the composite score. It was found that their mere presence may have adverse effects on responses to other items. The lack of a factor structure influenced by the negative item wording and the lack of a decrease in the magnitude of the validity coefficients due to negative item wording was surprising given previous findings in this area. It was suggested that existence of the adverse effects of negative items might be a function of some interaction between the characteristics of the items, respondents, and data collection procedures. Further study in this area was recommended to understand when negative items are expected to adversely affect questionnaire responses.
Subject Area
Quantitative psychology
Recommended Citation
Finney, Sara Jane, "A comparison of the psychometric properties of negatively and positively worded questionnaire items" (2001). ETD collection for University of Nebraska-Lincoln. AAI3009716.
https://digitalcommons.unl.edu/dissertations/AAI3009716