---
title: "Yapısal Eşitlik Modellemesi: Tanımlar ve Regresyondan Ayrılan Noktalar"
url: "https://merterbil.com/papers/yapisal-esitlik-modellemesi/"
author: "Mert Erbil"
pdf: "https://merterbil.com/papers/yapisal-esitlik-modellemesi.pdf"
---

# Yapısal Eşitlik Modellemesi: Tanımlar ve Regresyondan Ayrılan Noktalar

English title: Structural Equation Modeling: Definitions and How It Differs from Regression

A methods paper by Mert Erbil, from his Ph.D. studies, Advertising & Public Relations (dropped out), Bahçeşehir University, Istanbul. Written in Turkish. Cited 6 times on Google Scholar.

## Abstract

Social scientists spend a big part of their time doing the same thing, whatever their field. They build hypotheses, put them into a model and test them with numbers. For a long time regression was the default tool for that. But structural equation modeling (SEM) spread fast in the 2000s, and I wanted to understand why.

So the paper starts with the basics. It explains what SEM is, how it grew out of several multivariate analyses over the twentieth century and why it caught on with researchers from economics and marketing to advertising and sociology. Then it goes through its main features and the models you can build with it.

The second half is the comparison I was actually after. Regression explains how each independent variable affects the dependent one, which is fine for simple questions. The problem is, it can't describe the relationships between the independent variables on its own.

So testing a complex model with regression means running several separate analyses with different scales. SEM, on the other side, tests direct and indirect relationships in one model, path analysis included.

I wrote it in Turkish during my Ph.D. studies. It doesn't add a new method. But it puts the definitions, the history and the differences from regression in one place for anyone choosing between the two.

Full text (PDF): https://merterbil.com/papers/yapisal-esitlik-modellemesi.pdf

All papers: https://merterbil.com/papers/