Response styles such as acquiescence or extreme responding can bias correlations, factor structures, and prevent measurement invariance of personality inventories. Based on empirical data, we will discuss to what extent response styles are relevant, how response styles can be measured and controlled, and what the determinants of response styles are. Beatrice Rammstedt will illustrate that acquiescence biases the comparability of big five measures across countries. Julian Aichholzer and Meike Morren will introduce statistical models that allow controlling for acquiescence and extreme responding. Daniel Danner will demonstrate that acquiescence is not a general, uni-dimensional response style but that in it fact also depends on the item domain (such as personality or attitude items). Finally, Clemens Lechner will present data suggesting that acquiescence is not only associated with education but also with age related decline in cognitive functioning.

Effects of response set are often neglected in research investigating differences among groups. In contrast to individual diagnostics-which often controls for effects of social desirability for example personality assessments-investigating among group differences does not take into account effects of response styles. In this talk I will show that response style, and in particular acquiesce, have indeed strong biasing effects on personality assessments. In several different large scale population representative samples (n=888 to n=25,509) we proved that item responding is in particular in low educated groups strongly affected by acquiescence with blurring effects on the resulting factor structure. This effect could be shown to be generalizable across questionnaires, item formats, assessment modes, and numerous in particular Western countries. Implications of the findings for personality assessment are discussed.

Assessing measurement invariance (MI) is an important cornerstone in establishing equivalence of instruments and comparability of measured constructs. This study investigates how acquiescence response style (ARS) impacts the level of MI achieved (configural, metric, scalar). Data from a German representative sample (n = 3,118) were analyzed. The random intercept method is combined with multiple-group factor analysis to assess MI in a Big Five personality scale. Initial results suggest that if groups differ in ARS, neglecting that bias leads to different conclusions regarding the level of MI of the instrument. Implications and further applications are discussed.

Since the 1950s, extreme response style (ERS) has been associated with personality traits, such as anxiety, neuroticism, extraversion, conscientiousness, defensiveness, self-esteem, and depression. Inconsistent results have been obtained, for example, some find that extraversion relates positively to ERS, others find a negative relationship, and some find no relationship. This inconsistency might result from serious methodological challenges in exploring the relationship between response styles and personality. First, ERS can be measured by a sum score, a standard deviation score or a latent variable. Second, the modeling approaches to detect and correct for ERS diverges across studies. Third, most research uses the personality assessments both for measuring personality traits and response styles which inevitably leads to confounding style with content. Fourth, the personality assessments are affected by response styles themselves and need to be corrected. We propose a latent class factor approach that detects ERS using a validated scale and corrects for the influence of ERS on the personality assessments by simultaneously estimating the Big Five and ERS. Additionally, we assess the influence of methodological issues outlined above by comparing other modeling approaches to our model. We illustrate our approach using student data (n=200) from the Netherlands.

The present research investigates two facets of acquiescence: agreement and acceptance. Agreement has been defined as agreeing to all items (e.g., I am reserved, I am outgoing, I am not reserved; I am not outgoing) whereas acceptance has been defined as accepting opposite but non-negated items (e.g. I am reserved, I am outgoing) but not negated items (e.g. I am not reserved, I am not outgoing). Participants (n=398, 20-82 years old) completed a survey containing 96 items of different domains (personality, attitude, and knowledge items). The data were analyzed using hierarchical structural equation models. The results indicate that, (1) there is a general agreement factor that can explain about 2% of total item variance, (2) there are also domain-specific agreement factors that can explain up to 29% of total item variance, and (3) there is no general acceptance factor but domain specific acceptance factors that can explain up to 4% of total item variance. This suggests that acquiescence is not a general, uni-dimensional response style but has different facets which have different impact on items. Implications for research and assessments are discussed.

Acquiescence, or the tendency to respond to descriptions of conceptually distinct personality attributes with agreement/affirmation, constitutes a major challenge in the assessment of personality. The aim of this study was to shed light on cognitive ability as a potential source of individual differences in acquiescent responding. We hypothesized that respondents with lower cognitive ability exhibit stronger acquiescent response tendencies; this leads to problems in establishing the Big Five factor structure among these respondents, as opposed to respondents with higher cognitive ability. Further, we hypothesized that after controlling for acquiescence by using mean-corrected instead of raw item scores, the Big Five structure holds even at lower levels of cognitive ability. Analyses in a sample of 1,071 German adults aged 56 to 75 years using the Digit Symbol Substitution Test (DSST) as a measure of cognitive ability and the BFI-10, an abbreviated version of the Big Five Inventory, as a measure of personality, corroborated these hypotheses. This suggests that lower cognitive ability, and age-related declines in cognitive functioning more specifically, are associated with higher acquiescent responding in personality inventories; but that the problems this poses for establishing the five-factor structure can be resolved by statistically controlling for acquiescence.