本研究目的在多群組離散型因素分析模型下，發展一個檢測作答風格差
異的方法，共分為兩個子研究。研究一透過模擬資料操弄潛藏態度差異、四種常見作答風格類型與差異程度、樣本大小、問卷題數以及數種最小限制(minimum free-baseline, MFB) 六個自變項，比較具有作答風格差異的兩群組，在定錨題之外的各題各閾值的大小關係，最後再操弄閾值特徵符合程度之高低，使錯誤率在 5% 以下，作為檢定方法的依據。研究二分析「1998 年國際資訊科技教育應用研究 (SITES 1998)」之跨國五點量表問卷，選取其中六個國家，兩兩分析對於資訊融入教學的學習成效認同程度是否具有作答風格差異。
研究一結果顯示，隨著樣本數或作答風格差異程度增加，各種作答風格之閾值特徵皆會更加明顯，檢定的正確率也隨之增加，至於問卷題數則不影響檢定的正確率與錯誤率，但潛藏態度差異則會影響正確率。研究二結果顯示，雖然六國之間的各題各閾值幾乎都有顯著差異，但是在 30 個檢定中只有 11 個符合某種作答風格差異，這可能是國家間存在其他的作答風格差異，或是題目中還有其他的試題差異功能。而在作答風格顯示有差異的 11 個國家組合中，有6 個同時出現默認肯定與默認否定的作答風格差異，例如立陶宛對於挪威、法國與香港三個國家皆如此，可解釋為立陶宛相較於這三個國家，較不傾向回答中立選項，是一個值得注意的作答風格。
最後依據本研究結果可提出以下建議：在分析多群組態度量表時，先用各種 MFB 模型檢測群組間是否有作答風格差異，若沒有差異，則可以放心定錨一整題的閾值。若有差異，則根據作答風格差異的類型，選擇合適的 MFB 方式。例如群組間顯示出標準型的默認肯定作答風格，就選擇定錨第 1 個或第 2個閾值。

The purpose of this study is to develop a test for detecting the difference in response styles between groups under the multiple-group categorical confirmatory factor analysis model. In Study 1, the factors of impact of latent attitude, sample size, the number of questionnaire questions, and the type of minimum free-baseline (MFB) setting were manipulated and their effects on the empirical Type I error and power of the proposed test in detecting different types of response styles investigated. The results indicate that the greater the sample size or the larger degree of response styles, the greater the degree of compliance of thresholds’ characteristics and the power of the test. As for the number of questionnaire questions had no impact with type I & II error.
In Study 2, we analyzed the Likert scale data of Second Information Technology in Education Study 1998. In particular, six countries were chosen and examined for the presence of various response styles in their attitude towards the role of computer and other information and communication technologies. We found significant differences in response styles between 11 pairs of countries, among which 6 pairs showing the acquiescent and disacquiescent response styles simultaneously, such as Lithuania in comparison to Norway, France, and HongKong. The results implied that Lithuania was not inclined to midpoint response style with respect to the other three countries.
Based on our findings, some recommendations are given as follows: In the analysis of Likert scale, one can first start with a variety of MFB models to check whether there exists any response style between groups, if not, we can anchor one item as the common method to increase power. If so, we select the appropriate MFB method for the different types of response style. For example, while testing for the extreme response style between the two groups, we could choose to anchor the threshold of the middle category.

Asparouhov, T. & Muthén, B.O. (2006). Robust chi square difference testing with mean and variance adjusted test statistics. Retrieved from
https://www.statmodel.com/download/webnotes/webnote10.pdf

Chang, Y-W., Huang, W-K., & Tsai, R. (2015). DIF detection using multiple-group categorical CFA with minimum baseline approach. Journal of Educational
Measurement, 52, 181-199.

Drasgow, F.(1984).Scrutinizing psychological tests: Measurement equivalence and equivalent relations with external variables are the central issues. Psychological

Bulletin, 95, 134-135.
Greenleaf, E.A. (1992). Improving rating scale measures by detecting and correcting bias components in some response style. Journal of Marketing Research,26, 176-188.

Meisenberg, G., & Williams, A. (2008). Are acquiescent and extreme response styles related to low intelligence and education? Personality and Individual Differences, 44, 1539-1550.

Moors, G. (2004). Facts and artifacts in the comparison of attitudes among ethnic minorities: A multigroup latent class structure model with adjustment for response style behavior. European Sociological Review, 20, 303-320.

Muthén, B.O. (1983). Latent variable structural equation modeling with categorical data. Journal of Econometrics, 22, 43-65.

Muthén, B.O. (1984). A general structural equation model with dichotomous, ordered categorical, and continuous latent variable indicators. Psychometrika, 49, 115-132.

Muthén, B.O. (1998-2004). Mplus technical appendices. Los Angeles, CA:Muthén & Muthén.

Muthén, L.K. & Muthén, B.O. (1998-2015). Mplus User’s Guide. Seventh Edition. Los Angeles, CA: Muthén & Muthén.

R Core Team (2015). R: A language and environment for statistical computing.Vienna, Austria: R foundation for statistical computing. Retrieved from
http://www.R-project.org/

Satorra, A. (1999). Scaled and adjusted restricted tests in multi-sample analysis of moment structures. In Heijmans, D. D. H., Pollock, D. S. G. & Satorra, A. (Eds.), Innovations in multivariate statistical analysis: A festschrift for Heinz Neudecker (pp. 233-247), Dordrecht: Kluwer Academic.

Stark, S., Chernyshenko, O.S., & Drasgow, F. (2006). Detecting differential item functioning with confirmatory factor analysis and item response theory: Toward a unified strategy, Journal of Applied Psychology, 91, 1292-1306.

Willem, J.P., & Ronald E.A. (Eds.). (2001). ICT and the emerging paradigm for life-long learning. Amsterdam: IEA.

李盟、郭慶科 (2016)。反應風格與人格特質的關係。心理學進展，6(3)，320-331。