SİYASET BİLİMİNDE ETKİLEŞİMLİ MODELLER ÜZERİNE

Sosyal bilimlerde sayıları özellikle son yıllarda hızla artan çalışmada öne sürülen koşullu kuramsal beklenti ve önermelerin nicel yöntemlerle tahlil edilmesinde etkileşimli modellerden kullanılmaktadır. Ancak etkileşimli doğrusal ve doğrusal-olmayan modellerden yararlanan çalışmaların birçoğunda kurucu bağımsız değişkenlerinin koşulsuz, koşullu ve marjinal etkilerinin hesaplanması ve yorumlanması hatalı ya da eksik yapılmaktadır. Çalışmamızda siyaset metodolojisi alanında etkileşimli modellere dair güncel çalışmalarda dikkat çekilen hususlar ve önerilen yöntem ve tanı araçlarının detaylı bir incelemesi yapılmaktadır. Çok sayıda kuramsal ve kurgusal örnek yardımıyla, etkileşimli doğrusal ve doğrusal-olmayan modeller için önerilen bu yöntemlerin uygulanmasında dikkat edilmesi gereken noktaların altı çizilmekte ve bu noktalar çalışmanın sonuç bölümünde özetlenmektedir. Bu sayede ülkemizde gelecekte yapılacak çalışmalarda koşullu kuramsal beklentilerin öne sürülmesinde ve bunların nicel yöntemlerle tahlil edilmesi ve yorumlanmasında ortaya çıkabilecek olası hataların önüne geçilmesi amaçlanmaktadır.

Anahtar Kelimeler:

koşullu önermeler, etkileşimli doğrusal modeller, etkileşimli doğrusal-olmayan modeller, siyaset bilimi metodolojisi, nicel yöntemler

ON INTERACTIVE MODELS IN POLITICAL SCIENCE RESEARCH

Interactive models have been increasingly used in social science research to assess conditional theoretical expectations and hypotheses empirically. However, the direct, indirect, and marginal effects of the constitutive terms in interactive linear and non-linear models are often erroneously calculated and interpreted. This study examines the contemporary political methodology literature pointing to various sources of such errors and presenting alternative methods to calculate marginal effects and diagnostic tools. The proposed methods and statistical tools are then examined with the help of several applied and hypothetical examples, and several essential takeaways are summarized in the concluding section. Consequently, this study aims to help Turkish scholars avoid similar errors in positing conditional expectations, testing those using quantitative methods, and interpreting their findings.

Keywords:

Interactive linear models, interactive non-linear models, conditional hypotheses, political methodology, quantitative methods,

PDF

___

Achen, C. H. (1982). Interpreting and Using Regression. SAGE Publications.
Amorim Neto, O. ve Cox, G. W. (1997). "Electoral Institutions, Cleavage Structures, and the Number of Parties", American Journal of Political Science, 41/1, 149-174.
Berry, W. D., DeMeritt, J. H. R. ve Esarey, J. (2010). "Testing for Interaction in Binary Logit and Probit Models: Is a Product Term Essential?", American Journal of Political Science, 54/1, 248-266.
Bilecen, H. (2016). "Niche Party Success in Turkey: Do Policy Dimensions Matter?", Pamukkale Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 25, 139-157.
Brambor, T., Clark, W. R. ve Golder, M. (2006). "Understanding Interaction Models: Improving Empirical Analyses", Political Analysis, 14/1, 63-82.
Brambor, T., Clark, W. R. ve Golder, M. (2007). "Are African Party Systems Different?", Electoral Studies, 26/1, 315-323.
Clark, W. R. ve Golder, M. (2006). "Rehabilitating Duverger's Theory: Testing the Mechanical and Strategic Modifying Effects of Electoral Laws", Comparative Political Studies, 39/6, 679-708.
Cornell Statistical Consulting Unit. (2020). Overlapping Confidence Intervals and Statistical Significance. https://cscu.cornell.edu/wp-content/uploads/73_ci.pdf
Friedrich, R. J. (1982). "In Defense of Multiplicative Terms in Multiple Regression Equations", American Journal of Political Science, 26/4, 797-833.
Gelman, A. ve Pardoe, I. (2007). "Average Predictive Comparisons for Models with Nonlinearity, Interactions, and Variance Components", Sociological Methodology, 37/1, 23-51.
Hainmueller, J., Mummolo, J. ve Xu, Y. (2019). "How Much Should We Trust Estimates from Multiplicative Interaction Models? Simple Tools to Improve Empirical Practice", Political Analysis, 27/2, 163-192.
Hanmer, M. J. ve Kalkan, K. O. (2013). "Behind the Curve: Clarifying the Best Approach to Calculating Predicted Probabilities and Marginal Effects from Limited Dependent Variable Models," American Journal of Political Science, 57/1, 262-277.
Kemahlıoğlu, Ö. ve Sayarı, S. (2017). "Defecting Alone or Splitting Together? Individual and Collective Party Switching by Legislators", Public Choice, 171, 187-206.
Kentmen-Çin, Ç. (2015). "Participation in Social Protests: Comparing Turkey with EU Patterns", Southeast European and Black Sea Studies, 15/3, 223-237.
King, G., Tomz, M. ve Wittenberg, J. (2000). "Making the Most of Statistical Analyses: Improving Interpretation and Presentation", American Journal of Political Science, 44/2, 347-361.
King, G. ve Zeng, L. (2007). "When Can History Be Our Guide? The Pitfalls of Counterfactual Inference", International Studies Quarterly, 51/1, 183-210.
Moral, M. (2016). "The Passive-Aggressive Voter: The Calculus of Casting an Invalid Vote", Political Research Quarterly, 69/6, 732–745.
Rainey, C. (2016). "Compression and Conditional Effects: A Product Term Is Essential When Using Logistic Regression to Test for Interaction", Political Science Research and Methods, 4, 621-639.
Wright, G. C. (1976). "Linear Models for Evaluating Conditional Relationship", American Journal of Political Science, 20/2, 349-373.
Yıldırım, T. M. (2017). "Politics of Constituency Representation and Legislative Ambition under the Glare of Camera Lights," Legislative Studies Quarterly, 45/1, 101-130.
Zhirnov, A., Moral, M. ve Sedashov, E. (2022). Taking Distributions Seriously: On the Interpretation of the Estimates of Interactive Nonlinear Models. Political Analysis, First View, 1-22. https://doi.org/10.1017/pan.2022.9