Choosing the Optimal Number of Factors in Exploratory Factor Analysis: A Model Selection Perspective