多向度多層面羅西模式編製運動訓練反應量表

Development of Inventory of Sports Training Response Based on Multidimensional Many-facets Rasch Model

紀世清(Shyh-Ching Chi)

《大專體育學刊》 15卷4期 (2013/12) Pp. 421-428

https://doi.org/10.5297/ser.1504.004

摘要

本研究的目的是使用多向度多層面羅西模式編製運動訓練反應量表。本研究的受試者為598名選手，並以澳洲估計軟體程式分析資料。有關運動訓練反應量表的研究目前仍屬於單向度的基本假設，尚未發展多向度多層面Rasch模式。事實上，此運動教練評鑑量表包含七個因素且誤差變異來源涉及多個層面。對於「模式和資料的吻合的要求不理想，唯有使用多向度多層面方法編製運動教練評鑑量表，才會接近有效測量的理想。本研究結果發現：一、模式：以澳洲估計軟體估計多向度多層面的運動教練評鑑量表適合羅西測量模式。二、向度：就理論而言向度三「比賽」與向度四「醫學」是兩個不同的領域，在向度上要有區別，本資料為何會呈現高相關，有待進一步釐清。三、試題：多向度多層面評分量尺模式的試題參數估計結果：除第1－5題均方值沒有介於0.5－1.5範圍，其他題目均在均方值範圍內。評分量尺模式的試題參數估計信度得到0.95。以上資料大體顯示低分容易且高分難的順序，再者，各階段誤差很小，表示估計結果精準，足證明本研究的多重計分設訂合理。綜合上述的研究，本研究得到結論是：運動教練評鑑量表依據多向度多層面羅西模式編製是最合適的。

關鍵字

現代測量理論 ； 測量 ； 單向度 ； 總加評分量尺

並列摘要

The purpose of this study was to develop the inventory of sports training response based on the multidimensional many-facets Rasch model. Five hundred and ninety-eight athletes were participated in this study, and the collected data were analyzed by the ConQuest program. However, the past research was still considered as single dimensional basic hypothesis. The questionnaire includes 7 factors, and the error variations involve several dimensions, which was not ideal for model and data fit. We could achieve only effective evaluation through the application of the multidimensional many-facets model during questionnaire construction. The result showed that: (1) The Rasch model is fit for constraint estimation of multidimensional many-facets athlete instructs questionnaire. (2) The dimension, third factor ”competition” and fourth factor ”medical” are the two different domains, discriminate between two factors is necessary. However, the research shows a strong relation among the two factors, which needs further clarification. (3) The results of multidimensional many-facets summating rating scale showed that except questions 1,2,3,4, and 5, all other questions falls in between the MNSQ range of 0.5-1.5, with Separation Reliability 0.95, Chi-square test of parameter equality 8142.10, degrees of freedom at 60, and reached significant level The data showed that it was easy to have lower scores than high scores, and the error between each level was minimal. This revealed the accuracy of the study. In summary, the multidimensional many-facets Rasch model was found to be suitable for sports coach questionnaire.

並列關鍵字

multidimensional item response theory (MIRT) ； measurement ； unidimensional ； summating rating scale

參考文獻

Adams, R. J.,Wilson, M. R.,Wu, M. L.(1998).ACER ConQuest generalized item response modeling software.Melbourne, Australia:Australian Council for Educational Research.

Linacre, J. M.(2005).A user's guide to Winsteps ministep Rasch-model computer programs [Computer software and manual].Chicago, IL:Winsteps.

Spector, P. E.(1992).Summating rating scale construction.Newbury Park, CA:Sage.

Wright, B. D.,Masters, G. N.(1982).Rating scale analysis: Rasch measurement.Chicago, IL:MESA Press.

Wright, B. D.,Stone, M. H.(1979).Best test design: Rasch measurement.Chicago, IL:MESA Press.