Premium
A procedure for isolating social desirability variance in a three‐way component analysis
Author(s) -
LorenzoSeva Urbano,
Ferrando Pere J.
Publication year - 2012
Publication title -
british journal of mathematical and statistical psychology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 3.157
H-Index - 51
eISSN - 2044-8317
pISSN - 0007-1102
DOI - 10.1111/j.2044-8317.2011.02015.x
Subject(s) - variance (accounting) , set (abstract data type) , component (thermodynamics) , social desirability , preprocessor , data set , social desirability bias , computer science , matrix (chemical analysis) , response bias , core (optical fiber) , statistics , psychology , econometrics , social psychology , mathematics , artificial intelligence , accounting , telecommunications , physics , materials science , business , composite material , thermodynamics , programming language
Three‐way component analysis aims to summarize all the information in a three‐way data set related, for example, to the responses of individuals to a set of items in a set of situations. The data are decomposed into three component matrices (person, item and situation components) and a core matrix. Such data are generally obtained by means of a situation‐response questionnaire administered to a sample of participants. This paper proposes a preprocessing procedure for controlling social desirability in these kinds of questionnaires. The essential idea is to isolate the variance due to social desirability response bias and remove it from the data set. The procedure is demonstrated through its empirical application in the personality domain. Its performance is also assessed by means of a simulation study which shows that the presence of social desirability bias in a data set had different consequences: (1) the variance effect due to persons increased, and the effect of the triple interaction persons × items × situations decreased; (2) the person component and core matrix were partially distorted. However, the social desirability bias in the data set was successfully controlled by our preprocessing method.