DETECTING SOCIALLY DESIRABLE RESPONSES IN PERSONALITY INVENTORY
Loading...
Date
2011-07
Authors
GOVINDASAMY, PRIY ALATHA
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This study attempted to detect the socially desirable responses within differential
responses. Besides, this study also examined items and personality dimensions that are
vulnerable to social desirability. In the experimental design, a sample of 521 students
was tested twice with the International Personality Item Pool (IPIP) under honest and
socially desirable instructions. Responses from the frrst administration were classified as
honest group responses and those from the second administration were grouped as
socially desirable responses. The mean dispersion and fit analysis were applied in
detecting socially desirable responses. Differential Item Functioning (DIF) was used in
identifying items that are prone to social desirability followed by Differential Test
Functioning (DTF) wl-Jch examines the DIF effects to the test and to each personality
dimension assessed. Differential Step Functioning (DSF) was applied to determine the
contributing steps in polytomous responses to DIF. Factor analysis and item information
function curves were used to examine the personality dimensions that are susceptible to
social desirability. The socially desirable group reported a distinctively higher score
distribution at two standard deviations above the mean. Correspondingly, the socially
desirable group has a higher perc~ntage of non-fitting responses with values more than
2.0 logits. Differences of factor structures of IPIP dimensions were found between the
honest and socially desirable response groups. Test information function for the
Openness dimension illustrated a differential information function between the honest
and socially desirable groups. Six items were flagged for Differential Item Functioning
XI
(DIF) and three of the items represented the Openness dimension. In addition, the
presence of D IF caused changes in the test and in the Openness dimension for the
socially desirable groups. The DSF analysis reported response categories function
differently according to the items. The study concluded that high score and non-fitting
responses in a personality inventory are indications of socially desirable responding.
Meanwhile, low difficulty item parameter and endorsement of response categories are
also signs of social desirability. Therefore, screening of the potential socially desirable
responding individual and the elimination of items that are prone to social desirability
would help to arrest the problem of social desirability in high-stake testing.
Xll
Description
Keywords
DETECTING SOCIALLY DESIRABLE RESPONSES IN PERSONALITY INVENTORY