Course abstract
Self-report remains an important method for gathering personal data. However, scientific literature has amassed evidence of various sources of bias. This course will concentrate on three types of bias: acquiescence, social desirability, and overclaiming. Acquiescence refers to the tendency to endorse the highly positive categories on a scale, even for negatively worded items. This bias can result in overestimation of a general dimension and score reliability, among other issues. Managing acquiescence may involve the use of positively and negatively worded items and employing techniques in the framework of CTT or IRT. Social desirability involves the inclination to endorse items that create a favorable impression. This phenomenon arises from the interaction of item content, personal characteristics, and assessment context. Strategies to estimate social desirability include specific scales, but these may introduce additional bias due to their correlation with measurement dimensions. An alternative approach is to manipulate item content by rephrasing and balancing desirable and undesirable aspects. A second step involves modeling the manipulated items using bifactor models. Overclaiming is the tendency to positively endorse items describing personal competencies or knowledge, even when the respondent lacks those skills. This reflects a form of cognitive faking. Estimating overclaiming can be achieved by employing 'foil' items, such as asking if the respondent is familiar with Mozart's Ninth Symphony. The overclaiming scores are then estimated using CTT or IRT. The course will provide theoretical frameworks for understanding acquiescence, social desirability, and overclaiming, and it will offer practical examples for modeling each of these biases.