Response tendency

from Wikipedia, the free encyclopedia

A response tendency (engl. Response set , response bias ) response distortion or response error is a systematic deviation of the in surveys , interviews u, opinion polls. a. social science surveys , psychological tests and questionnaires received responses. The data then does not form the applicable (actual and "true") settingsand facts. The reasons for such deviations are to be found in the respondents, in the questions and in the design of the questionnaire, in the particular situation or in the interviewer ( interviewer effect ).

Response behavior

Every survey by an interviewer or a questionnaire represents a social situation that can have an influence on the answers. The investigators can also inadvertently influence the responses and other behaviors through their presence and their expectations. Above all, however, the individual attitudes of the respondents and their personality traits will have an impact, because every self-assessment is also a self-portrayal . Therefore, pronounced response tendencies can also be viewed as important facets of personality traits .

The bias due to response tendencies is a typical method problem for questionnaires. But such effects will undoubtedly also appear in interviews and in most other psychological diagnostic procedures. The effects of methodological reactivity , i.e. H. the influence of the examination method used on the results is a fundamental problem of almost every psychological methodology, even many medical examination methods. Psychological and medical examinations can also lead to the phenomenon of reactance , i. H. to defensive reactions and refusal to participate.

Response tendencies were examined from various professional perspectives, including: a. test methodological, differential psychological, social psychological, cognitive psychological. This topic also includes schematic judgment tendencies, characteristic judgment errors and systematic memory delusions. A related topic is formed by the phenomena known from medical diagnostics, especially in psychiatry, of simulation, aggravation , dissimulation or trivialization of complaints and findings; the frequent discrepancies between the physical symptoms experienced and the objective findings form an important issue in the psychology of behavioral illness .

Most of these response biases and judgment errors can be found as early as the 1950s in Lee J. Cronbach and Joy P. Guilford, et al. a. the tendency to say yes (acquisition) and the tendency towards the undecided middle or towards the ends of a scale (extrema).

Typical response bias

Several effects are considered cause for the unit variance methods (Engl. Common-method variance ).

Method effects

Respondent characteristics

A distinction is made between formal answer tendencies, which prefer certain answers regardless of the content of the question, and content-related answer tendencies, which are influenced by the content and sequence of the question.

Formal response tendencies

  • Acquiescence : The yes-saying tendency (content-independent tendency to agree, acquiescence tendency) is the tendency of people to answer questions with "yes", "true" or "right", regardless of the content of the questions. The yes-saying tendency is found more often in authoritarian personalities and fearful and conservative people than adapted behavior.
  • Trend towards the middle : The tendency towards the middle (error of central tendency), also extremely shy, is the tendency of those surveyed to select the middle scale points for multi-level scales (e.g. Likert scales ).
  • Tendency towards mildness / hardness or towards extreme judgments: The tendency towards mildness or hardness (error of extreme tendency) is the tendency of respondents to tend to extremes with multi-level answer options. This is especially true in exam or test situations.
  • Position effects: depending on the position of the question, different answers can be given - while answering the question, the understanding of the instruction can change or, at the end, satiety and fatigue effects can influence the answers.
  • Memories are also affected by the prevalence error (or base rate error). Here, decisions are made on the basis of a statistical relationship that is present in the memory, without going into the information available. (Example: Linda problem; see representativity heuristic )

Content-related response tendencies

  • Social desirability ( social desirability response set) results from the tendency to answer items not according to the personally applicable attitude, but according to social norms that the test subject believes are desirable. The tendency towards social desirability is mostly seen as a widespread, more or less unintentional tendency towards positive self-expression. Methods to increase the number of honest answers through anonymization are the randomized response technique and the unmatched count technique (see also sensitive question ).
  • Consistency bias results from the tendency to answer similar-sounding statements coherently, so that they fit together in terms of content, even if this is not the case.
  • Retrospection effect (recall bias, retrospection effect) means that experiences and events in retrospect, on the next day or after a few weeks are assessed more positively or negatively, e.g. B. the pain experienced in the memory more intense than in the currently collected assessments (negative retrospection effect).
  • Rezenzeffekt (Engl. Recency effect) states that later incoming information a greater impact on the memory capacity of a person to exercise than before incoming information. The short-term memory gives the most recently perceived information more weight (see primacy recency effect ).
  • In contrast to the rezen effect, there is the primary effect . It is assumed here that information received at the beginning can be better stored in the long-term memory and can thus be remembered more easily. (see primacy recency effect )
  • Hindsight bias (Engl. Hindsight bias) describes inaccurate memories when people after they have experienced the actual outcome of an event, remember systematically wrong with their former own predictions.
  • Distortion of silence ( non-response bias ): Respondents may respond differently to those who did not respond if they had participated in the survey. The missing answers thus distort the overall picture.
  • When using interview methods, results can be distorted by the interviewer through conscious or unconscious influence by the interviewer, which can be counteracted by standardization.

Influences of formulation and design

The answers can be influenced by the formulation of the questions (wording) and (in the case of questionnaires also the given answer options), by the comprehensibility of the questions and the instructions as well as by the course of the investigation. This includes unnecessary foreign words and ambiguous formulations, or those that already suggest a certain answer ( framing effect ). The order of the questions can also cause distortions. Order effect (also: question series effect or position effect ) if a previously asked question influences how the content of the following question is understood and evaluated.

Anyone developing a questionnaire should refer to the relevant textbooks and carefully check the design before use (see questionnaire , questioning technique ).

Influences of the examiner and the examination situation

Many of the aforementioned effects on the respondent's (examined) side have similarities on the investigator's (interviewer) side, e.g. B. Observation and assessment errors as a halo effect or certain characteristics of the interviewer ( result distortion by the interviewer ). If the examiner himself reveals certain expectations, an experimenter effect occurs (also: Rosenthal effect , Pygmalion effect , Rosenhan experiment or experimenter artifact). The Hawthorne effect describes that participation in an investigation can trigger special expectations that lead to bias of the results.

criticism

Conceptually, different response tendencies are distinguished, but these can hardly be distinguished from one another empirically and methodologically. They are often linked to one another and are influenced by the individual cognitive style, semantic and linguistic difficulties, the sequence effects, etc. In addition, many of these tendencies, such as the tendency to say yes, the tendency towards social desirability, or the tendency towards extreme or intermediate answer categories, are typical characteristics of certain personality traits.

In the test methodology, the opinion was widespread at times that these effects were controlled or balanced: by pairs of oppositely polarized questions (which is often linguistically complicated), by specifying two-stage answer options (only “yes” or “no”), by control questions or by a so-called lie scale, which should capture contradictions and probably incorrect answers. The attempt to separate conscious and unconscious parts or the desire for a test method isolation and statistical correction generally overwhelm the questionnaire method. Clarifying how a person “actually” is reminiscent of an earlier and outdated conception of an immutable personality and does not correspond to the understanding of personality traits with time and situation-dependent variability . In particular, the investigation of social desirability requires an interactionist perspective (in the sense of interaction ) that establishes a connection: between test situation, test motivation, personality traits , stylistic features, verbal intelligence , personal expectations and motives, willingness to adapt, weighing up benefits, fear of disadvantages or fear of being Discovery as well as the individual and general dispositions of values ​​such as honesty and openness. There will also be a wide range of opinions as to which answers are desirable or not in a particular situation or task.

The self-presentation is an integral part of the self-assessments expressed. Anyone who uses self-assessments in psychological diagnostics must also accept the structural subjectivity of these self-reports. Trying to delimit individual facets of self-portrayal within a person's self-assessment seems to be only interpretatively possible, unless objective life history data or behavioral observations can be included. In the social psychological attitude research reported objectification attempts by methods of psychophysiology and the Bogus pipeline technique with the deception about an alleged lie detector are unsuitable for different professional ethical or methodical reasons.

The guidelines for quality assurance in psychological diagnostics require that the results of the examination can be falsified as little as possible by the candidate himself. So far, however, there are no rules or conventions for the evaluation of questionnaire methods and interviews as to how this can be reliably recognized or prevented.

Individual evidence

  1. ↑ Lee J. Cronbach: Essentials of psychological testing . 3rd. ed. Harper and Row, New York 1970.
  2. ↑ Joy P. Guilford: Personality. McGraw Hill, New York 1959.
  3. ↑ Williams, LJ, Hartman, N. & Cavazotte, F. (2010): Method Variance and Marker Variables: A Review and Comprehensive CFA Marker Technique. Organizational Research Methods, Volume 13, No. 3, pp. 477-514.
  4. ^ Ian Needham: Nursing Planning in Psychiatry . Recom, 3rd edition 1996, page 73. ISBN 978-3-89752-034-9 .
  5. ^ Del Paulhus: Social desirable responding: The evolution of a construct. In HI Braun, DN Jackson, DE Wiley (Eds.). The role of constructs in psychological and educational measurement. Erlbaum, Mahwah, NJ. 2002, ISBN 0-8058-3798-1 , pp. 49-69.
  6. ↑ see Mummendey, 1995

literature

See also