Premium
A method for the investigation of irrelevant response set in ordered metric and original questionnaires
Author(s) -
Phillips J. P. N.
Publication year - 1979
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.2044-8317.1979.tb00597.x
Subject(s) - dependency (uml) , metric (unit) , sequence (biology) , set (abstract data type) , preference , statement (logic) , sample (material) , statistics , test (biology) , mathematics , psychology , computer science , artificial intelligence , operations management , paleontology , chemistry , chromatography , biology , political science , law , economics , genetics , programming language
A statistical method is described for the analysis of response sot in ordinal questionnaires (in which a sequence of statements is presented to the patient, who is asked to say to each whether he feels ‘better or worse than it says on the card’) and ordered metric questionnaires (in which a sequence of pairs of statements is presented, and the patient is asked to say whether the ‘Top’ or the ‘Bottom’ statement better describes how he feels). In the first case, sequential dependency in the sequence of ‘hotter’ and ‘worse’ responses can be tested by means of the Wald‐Wolfowitz runs test. In the second, there are three types of response set to be tested: bias, or preference for ‘Top’ or for ‘Bottom’ responses irrespective of content; sequential dependency in the sequence of implied Health and Illness responses; and sequential dependency in the sequence of ‘Top’ and ‘Bottom’ responses. Significance tests for indices of each are given: the indices themselves are in general correlated (and a large sample method of taking account of the correlations is given), but it is described how to arrange the questionnaire so as to make them as uncorrected as possible. A suite of computer programs for preparation of questionnaire material and scoring and analysis of responses is reported. Possible generalizations of the method are mentioned.