Further Studies of Test-Retest Effect on Personality Questionnaires

Windle, Charles

doi:10.1177/001316445501500304

Cited by 64 publications

(44 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This intriguing mean answer shift on retest continues to be observed for various tests (Chance, 1955;Goldberg, 1978;Payne, 1974;Perkins & Goldberg, 1964;Windle, 1955). We find that the increased scores on retest are not really a test-retest phenomenon but an item-to-item reaction that is also apparent within the first test.…”

Section: Test-retest Reliability Shiftssupporting

confidence: 63%

“…Since then many researchers have found that people answer a personality retest more consistently than they answered the first test (D. W. Fiske, 1957;Goldberg, 1978;Howard, 1964;Howard & Diescnhaus, 1965;Schubert & Fiske, 1973;Windle, 1955). Many test developers, including Taylor (1953) for her anxiety measure, report that retests have higher internal consistency than first tests.…”

Section: Test-retest Reliability Shiftsmentioning

confidence: 99%

See 1 more Smart Citation

Context Effects in Social and Psychological Research

Schwarz¹,

Sudman²

1992

209

View full text Add to dashboard Cite

), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use of general descriptive names, trade names, trademarks, etc., in this publication, even if the former are not especially identified, is not to be taken as a sign that such names, as understood by the Trade Marks and Merchandise Marks Act, may accordingly be used freely by anyone.Camera-ready copy supplied by the authors. Printed and bound by Braun-Brumfield, Inc., Ann Arbor, MI. Bibb Latane and Deborah Richardson hosted the conference at the Nags Head Conference Center in Kill Devil Hills, North Carolina. Their hospitality, the joys of southern cooking, and the outer banks of North Carolina greatly contributed to an enjoyable and productive meeting. Special thanks are due to Mary A. Spaeth, who edited all contributions to this volume and turned a pile of manuscripts into a readable book. Finally, we thank all conference participants for lively and controversial discussions from breakfast to midnight and for the fine chapters that they contributed.

show abstract

Section: Test-retest Reliability Shiftssupporting

confidence: 63%

Section: Test-retest Reliability Shiftsmentioning

confidence: 99%

Context Effects in Social and Psychological Research

Schwarz¹,

Sudman²

1992

209

View full text Add to dashboard Cite

show abstract

“…Similarly, on personality inventories, subjects, as they respond to the items, may realize that the items vary -18-systematically in social desirability and proceed to respond in terms of this variable. A somewhat different explanation of these trends is that they are caused by changes in the subjects' approach to the test (Caldwell, 1959;Fiske, 1957;Fiske, 1961;Mccreary & Bendig, 1954;MCGeoch &Whitely, 1927;MolJenkopf, 1950;Voas, 1956;Windle, 1955)--the energy, interest, and anxiety elicited by the test. For example, subjects' energy or interest may decrease as they respond to the test items, and they may then tend to make responses consistent with their stylistic predilections, which requires less expenditure of energy than pondering the items before making responses to their content (Jackson, 1959;Stricker, 1963).…”

Section: Discussionmentioning

confidence: 99%

“…Such research on changes between two administrations of the same test has already begun (Caldwell, 1959;Windle, 1955), but a similar investigation of trends within tests is still lacking.…”

Section: Discussionmentioning

confidence: 99%

Difficulty and Other Correlates of Criticalness Response Style at the Item Level1

Stricker¹

1964

ETS Research Bulletin Series

View full text Add to dashboard Cite

This study examined the.relationship between the difficulty of items on an achievement-type test of report writing and (a) the items' correlations with the test's criticalness response style score, and (b) the items' correlations with the test's content score. Item readability and item format variables were also analyzed. The major findings were that (a) the items' difficulty was related to their correlations with the content score and unrelated to the items' correlations with the response style score; (b) the items' location in the test was generally related to their correlations with the response style score, and unrelated to the other variables; and (c) the readability indexes were not related to the three basic item variables. DIFFICULTY AND OTHER CORRELATES OF CRITICALNESS RESPONSE STYLE AT THE ITEM LEVEL 1Largely as a result of Cronbach's (1946; reviews of the early literature on response sets--or response styles--on tests, the view is widely held that tests and test items that are difficult or ambiguous are most affected by those response styles, such as acquiescence, evasiveness, and extremeness, that are linked to the response format.Despite the prevalence of this view, the results of relevant studies are not altogether convincing, and are largely limited to two response styles--acquiescence, and the position of the chosen alternative on multiple-choice tests.One major group of relevant studies examined acquiescence on tests or items of varying difficulty or ambiguity. Only one of the studies (Gage, Leavitt, & Stone, 1957) concerns the difficulty issue. In a comparison of difficult and easy general information items, the reliability of the number of true responses was .68 for the 50 difficult items, and .09 for the 40 easy ones. The two kinds of items, however, did not represent exactly the same content areas.The other studies of this kind concern various facets of item ambiguity. Two of these studies directly concern ambiguity. In one study (Bass, 1955), the original and reversed version of the California F scale items that were rated most opposite in meaning (and, hence, assumed to be clear in meaning) appeared to measure content (i.e., the correlation between the two items was relatively high and in the content direction), while those rated less opposite in meaning (and, hence, assumed to be -2-relatively ambiguous in meaning) did not, evidently because of the operation of acquiescence on the latter items. Moreover, an analysis of variance found significantly (£ < .01) greater acquiescence to the pairs of items that were least opposite in meaning. The assumption in this study that the rated level of opposition measures ambiguity is questionable, for it is possible for a pair of items that are rated as similar in meaning to be very upambiguous. In a second study (l3a.nta, 1961), the number of agree responses on attitude scales increased and the number of extreme responses decreased with the ambiguity of the scales' referents. The referents were "President Eisenhower," "College Fraternities,"...

show abstract

“…The reference group design allows to compare the 2 treatments against the control group (WG) with high external validity because external events and retest effects are parallel. (Retest effects [Windle, 1954[Windle, , 1958 usually go in the direction of 'mental ameliora tion '. ) However, the latter becomes false when we include follow-up measure ments.…”

Section: Design and Statistical Methodsmentioning

confidence: 99%