2006 Fiscal Year Final Research Report Summary
A Study of the Impact of Measurement on the Results of Subsequent Measurement and Its Implications in Quantitative Social Research.
| Project/Area Number |
17530397
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Sociology
|
|---|
| Research Institution | BUKKYO UNIVERSITY |
|---|
Principal Investigator |
YAMAGICHI Yoh BUKKYO UNIVERSITY, Faculty of Sociology, Associate Professor, 社会学部, 准教授 (00262543)
|
|---|
| Project Period (FY) |
2005 – 2006
|
| Keywords | social research method / sociology / cognitive science / reactivity / reliability / validity |
|---|
| Research Abstract |
This study examines reactivity of measurement and its implications through reviewing existent literature and analyzing experimental data. This study focuses on the impacts of reactivity on reliability and validity assessment (Chapter 1).To examine the impacts, the change of correlation among variables by carry-over effect is discussed in Chapter 2. Grasping this phenomenon involves methodological difficulties, but it may have great consequences for reliability and validity assessment. Chapter 3 examines the consequences of reactivity on internal-consistency in personality-scales. Knowles's studies and the follow-up examinations of the present study demonstrate correlations between item reliabilities (item-total correlations) and the serial positions of every item in several personality tests (single trait). The present study discusses self-generated internal-consistency. In other words, the scales with a larger number of items tend to show higher item reliability because of higher correlations among items (not because of compensation for random errors with each other). Chapter 4 examines the impact of reactivity on predictive validity. Especially, the predictive validity of measurement of intention is examined in the case the criteria concerned actual behaviors. Reconsideration of Sherman's insight about the self-erasing nature of errors of prediction demonstrates self-generated predictive validity. The conditions for occurrence of this phenomenon are examined using the results of experiments of the present study. Chapter 5(a supplement) highlights the role of buffer items that have been used to reduce carry-over effects. The last chapter discusses the self-contradictions in many of the reliability and validity assessment methods. In other words, while those assessment methods involve the procedures of multiple measuring, they neglect reactivity that is the characteristic of the procedures.
|
