Investigating the Performance of the Exploratory Graph Analysis When the Data Are Unidimensional and Polytomous

The question of how observable variables should be associated with latent structures has been at the center area psychometrics. A recently proposed alternative model to traditional factor retention methods is called Exploratory Graph Analysis (EGA). This method belongs broader family network psychometrics which assumes that associations between observed are caused by a system in have direct and potentially causal interaction. approaches psychological data an exploratory manner enables visualization relationships allocation dimensions deterministic manner. In this regard, aim study was set as comparing EGA when unidimensional items constructed polytomous response format. For investigation, simulated sets were used three different conditions manipulated: sample size (250, 500, 1000 3000), number (5, 10, 20) internal consistency scale (α = 0.7 α 0.9). results revealed robust especially graphical least absolute shrinkage selection operator (GLASSO) algorithm provides better performance true dimension than Kaiser's rule yields comparable other (optimal coordinates, acceleration Horn's parallel analysis) under some conditions. These discussed based on existing literature suggestions given for future studies.