項目滿意度之重要性加權：以生活品質測量為例

Abstract

在生活品質研究中，許多生活品質測驗在計算個人的生活品質分數時，採用重要度加權的方法，將個人在各項目所評量的滿意度分數與重要度分數相乘，得出重要性加權的生活滿意度分數，並依此作為個人生活品質的指標。然而，根據Locke的情感間距假設（range-of-affect hypothesis, 1969, 1976），當個人在進行項目滿意度評量時，該項目的重要程度已經介入滿意度評量的歷程，而影響滿意度評量的強度。因此，本論文以Locke的情感間距假設為架構，以生活品質測驗為範疇，探討項目滿意度分數之重要度加權的適當性。本論文共有三個次研究，研究一旨在比較未加權滿意度分數與加權後滿意度分數，對整體生活滿意度的預測效果是否有所差異。研究二則以問卷調查法檢驗Locke的假設，了解在生活品質測量中，項目重要度是否影響項目滿意度的強度。研究三則是採用情境故事的實驗法，檢驗Locke的假設。研究結果發現，加權分數與未加權分數對整體生活滿意的預測效果並無顯著差異，且項目滿意度對於整體生活滿意度的預測力，並不會因為項目的重要程度而有所差異，顯示重要度加權無助於整體生活滿意度的預測。此外，研究結果也發現項目重要度對項目滿意度的強度具有影響力，顯示項目滿意度評量已涵蓋項目重要度的訊息。因此，本研究認為項目滿意度之重要性加權並不是一個有效且必要的計分程序。

“Importance weighting” is a common idea in quality of life (QOL) measurement. Based on the widespread idea that important domains should have more contribution to individuals’ QOL total score, the weighting procedure of multiplying item satisfaction by item importance has been adopted in many QOL instruments. However, in Locke’s (1969, 1976) range-of-affect hypothesis, he indicated that the satisfaction evaluation of an item was determined by the have–want discrepancy, by the importance of the item, and by their interaction (discrepancy × importance), implying that item satisfaction has incorporated the judgment of item importance and that weighting item satisfaction score with item importance score is redundant. Several studies on job satisfaction have supported Locke’s range-of-affect hypothesis; however, the range-of-affect hypothesis has not been examined in QOL research, and the conclusion derived from the hypothesis might not be directly applied to QOL research. Therefore, the purpose of this thesis is to address the issue of importance weighting by examining Locke’s hypothesis in the context of QOL research. Three studies are conducted to address this issue. In order to understand the empirical effect of importance weighting, in the first study, several weighting algorithms proposed by different scholars are adopted to investigate whether the weighted satisfaction scores would have better predictions of a criterion variable (a global life satisfaction measure) than the unweighted satisfaction scores. The second study is conducted to examine the range-of-affect hypothesis by a survey study for testing the hypothesis in the interpersonal context. In this study, moderated regression analysis is conducted to test whether the association between item discrepancy and item satisfaction is stronger among people who regard the item as having higher importance than people who give it lower importance. Finally, in the third study, a within-subject experiment is conducted to test the range-of-affect hypothesis in an intrapersonal context. This study aims to show that, given an amount of discrepancy, people have stronger satisfaction/dissatisfaction on a higher-importance life aspect than on a lower-importance aspect. Generally, the results of the three studies are consistent with the previous studies on job satisfaction, revealing that (1) weighing item satisfaction with item importance did not contribute to predicting a criterion variable and (2) item satisfaction has incorporated the judgment of item importance. Based on these results, the importance weighting procedure is unnecessary. In addition, several points derived from the issue of importance weighting are also discussed, such as the value of importance rating on QOL measurement, the psychological mechanism on satisfaction response, the relationships between domain satisfaction and global satisfaction, and the implications for QOL promotion. From these discussions, it is obvious that QOL research is a broad field that relates to various issues. Moreover, to thoroughly understand individuals’ QOL, there are many issues that remain to be further explored.