Academic journal article Psychologische Beiträge

Towards the Identification of Non-Scalable Personality Questionnaire Respondents: Taking Response Time into Account

Academic journal article Psychologische Beiträge

Towards the Identification of Non-Scalable Personality Questionnaire Respondents: Taking Response Time into Account

Article excerpt

Summary

Starting with an overview of the literature relevant for this topic, it is depicted that within the framework of research in the meaning of response latencies for completing personality questionnaires, the approach of the Item-Response-Theory has not been adopted yet. Thus, the study on hand deals with the attempt to identify so-called non-scalable subjects by means of a mixed-Rasch-model and to clarify the meaning of the response latencies for this group. Data of the Eysenck-Personality-Profiler 'German-version' (EPP-D) reveal that response latencies are an important indicator to verify whether the rating-scale in the questionnaire has been used correctly by the subjects. With respect to the EPP-D, this is the case for just a small percentage of subjects. The necessity is pointed out for taking into account the test theory of the questionnaires used for the analysis of the meaning of response latencies.

Key words: response latencies, personality questionnaire, Eysenck-Personality-Profiler

Introduction

Recently there has been renewed interest in the response latencies of personality questionnaires. However the psychological meaning of response latencies is quite unclear. Since there are different approaches to using information of response latencies in personality questionnaires, it is worth having a systematic overview.

1. Studies on the relationship between response latencies and items' attributes

It seems obvious that latencies can tell us something about the item itself. According to the early studies by Dunn, Lushene and O'Neil (1972), Rogers (1973) and Stout (1981), response latencies are sensitive indicators of item characteristics, such as length, ambiguity or social desirability.

Amelang, Breit and Schneyer (1985) found that response latencies observed with items of the Freiburger Personality Inventory (FPI) (Fahrenberg, Selg & Hampel, 1978) do not only depend on the length of an item and on the equivalence of the distractors, but also on the content of an item: Answering items with more trait-specific content required more time than answering items with state-specific content. An additional finding was that extreme ratings coincide with shorter reaction time than mean ratings.

A follow-up study (Amelang, Eisenhut & Rindermann, 1991) reports similar results for an adjective checklist. The authors of the study defined an answering model differentiating between state and trait ratings to predict reaction time, which holds for adjectives suitable for describing traits and states. According to this model, situation sampling necessary for trait-- ratings requires more time than indicator sampling when answering on state-ratings. Again, extreme ratings coincide with shorter reaction time than mean ratings.

Another study with an adjective checklist (Amelang, 1994) reports no relevant impact of the adjective's broadness on response latencies. The findings of the 1991 study - longer reaction time for trait-ratings - could only be replicated for adjectives not referring to socially desirable attributes.

A recent study by Kinney (1999) reports that item clarity and agreement probability did not relate to response latencies on standard test items of 16PF (Cattell, Eber & Tatsuoka, 1970).

2. Studies on the relationship between response latencies and subjects' attributes

A lot of studies favor the general assumption that response latencies indicate the response quality. While some studies report how response latencies affect reliability and validity of a scale, believing in latencies as an indicator for e.g. self-concept or response styles such as honesty, other studies focus on the relation of response latencies and moderating variables such as impression management (faking). Neubauer and Malle (1997) classify studies on the relation of response latencies and response quality in studies using indirect investigation methods on the one hand and in studies using direct investigation methods on the other. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.