The Value of Narrow Characteristics or Components

In addition to efforts to understand the structure of behavior and personality, some researchers have  expressed concerns about the broad nature of the domains included in these models and have  argued for a focus on narrower components (i.e., more specific aspects of behavior within each  broad domain). Research suggests that component-level domains (e.g., planning) may result in better prediction of narrower, more specific outcomes (Ashton et al., 1995; Hogan & Roberts, 1996; Mount & Barrick, 1995; Ones & Viswesvaran, 1996; Schneider, Hough, & Dunnette, 1996). Conversely, this basic argument suggests that broad, global domains should better predict broad, global criteria than narrow, component-level characteristics (Cronbach, 1960). For example, Stewart (1999) showed that while global Conscientiousness was related to all stages of job performance, components such as Order were more strongly associated with early stages of job performance, whereas Achievement was associated with later maintenance stages of performance. Research shows that some components influence behavioral outcomes more strongly than global characteristics (e.g., Schouwenburg & Lay, 1995). For example, a number of studies have found that the Conscientiousness components Achievement Striving, Dependability, and Dutifulness have validity over and above a general measure of Conscientiousness in predicting academic performance outcomes (Chamorro-Premuzic & Furnham, 2003; De Fruyt & Mervielde, 1996; Gray & Watson, 2002; Lievens, Coetsier, De Fruyt, & De Maeseneer, 2002). Importantly, the value of components as predictors has been explicitly addressed in the HEXACO model, with each factor being further subdivided into components (Ashton & Lee, 2009; Ashton, Lee, & de Vries, 2014). Given the present state of the literature, it is clear that including component-level information in a hierarchical framework to organize and describe the behaviors important for predicting a range of education and work outcomes is an ideal starting point.