Kibeom Lee, Ph.D., & Michael C. Ashton, Ph.D.


The Early HEXACO Personality Inventory

We began construction of the HEXACO Personality Inventory in the year 2000. Our aim was to assess the six personality dimensions found in lexical studies of personality structure as conducted in various languages, and also to reflect our theoretical interpretations of those factors (see, e.g., Ashton & Lee, 2001, 2007). By 2002, we were using a provisional 108-item self-report version of this inventory in our research. The six scales each contained 18 items that spanned a wide array of content, but were not divided into "facet"-level subscales assessing distinct narrow traits.

Inclusion of Facet-Level Scales

The next step was to generate additional items that would assess distinct traits within each factor. We developed definitions for those facet scales with the aim of representing a wide array of content within each factor. The resulting version of the HEXACO Personality Inventory consisted of six broad factor scales, each subsuming four narrower facet scales (Lee & Ashton, 2004). In addition to producing the self-report form of the inventory, we also adapted the items for an observer report form. The full-length version consisted of 192 items (i.e., eight items per facet scale), and the half-length version consisted of 96 items (i.e., four items per facet scale).

Interstitial Facet Scales

We next added two more facet scales to the HEXACO-PI (see Lee & Ashton, 2006). Unlike the original 24 facets, each of which was assigned univocally to one of the six broad factors, these two new facets were "interstitial" scales intended to assess some important and interesting traits that load moderately on two or more of the six factors. The inclusion of the two new interstitial scales increased the length of the HEXACO-PI from 192 to 208 items in the full-length version, and from 96 to 104 items in the half-length version.

One of these interstitial facet scales was Altruism versus Antagonism. This facet was developed to assess a trait of sympathy and softheartedness. In lexical studies of personality structure, these traits tend to "migrate" between the Honesty-Humility, Agreeableness, and Emotionality factors, a finding that is consistent with our theoretical interpretation of the factors (Ashton & Lee, 2001, 2007). The inclusion of this facet scale allowed us to capture the specific variance associated with this trait, and thus to allow better prediction of various altruistic (versus antagonistic) behaviors that are of obvious importance in human interactions.

The other interstitial facet scale was Negative Self-Evaluation, which was intended to assess a construct of very low self-esteem. This facet, which loaded negatively on Extraversion and positively on Emotionality, was added partly because of the unique importance of self-evaluation as a personality variable, and partly because of the relevance of negative self-evaluation to depression and to various personality disorders.

The HEXACO Personality Inventory-Revised

More recently, we modified the HEXACO-PI to produce the HEXACO Personality Inventory-Revised (HEXACO-PI-R). Two changes differentiate the HEXACO-PI-R from the previous HEXACO-PI. First, one of the facet scales within the Extraversion factor was replaced: specifically, the HEXACO-PI Expressiveness facet has been deleted, and a new facet called Social Self-Esteem has been added. Second, the interstitial scale of Negative Self-Evaluation, described above, has also been removed.

The removal of the Expressiveness facet was in part due to some difficulties associated with this facet in translated versions of the HEXACO-PI. To begin, translators of the inventory found the items of this scale to be difficult to translate from English. Moreover, this facet scale tended not to show strong loadings on the Extraversion factor in some translated versions of the HEXACO-PI, and also tended not to be strongly correlated with markers of the lexical Extraversion factor.

In considering these results, we decided that the Extraversion factor would be better defined by a construct of social self-esteem than by that of expressiveness. Our social self-esteem construct overlaps with self-esteem as traditionally conceptualized, in the sense of being defined by self-evaluation. However, social self-esteem emphasizes self-esteem in the interpersonal context, emphasizing one's sense of being popular and liked by others.

In adopting the new Social Self-Esteem facet, we also decided to remove the Negative Self-Evaluation facet. The latter scale possessed a strongly skewed distribution, and we judged that nearly all of the predictive validity of Negative Self-Evaluation would be captured by the Social Self-Esteem facet in combination with other HEXACO-PI facets (e.g., the Anxiety facet of the Emotionality factor).

Thus, the HEXACO-PI-R assesses four facet-level scales within each of the six factors; in addition to these 24 facets, the interstitial Altruism versus Antagonism variable is a 25th facet scale. The HEXACO-PI-R therefore contains 200 items in its full-length version (HEXACO-200), or 100 items in its half-length version (HEXACO-100).

Note that, when calculating scores for the six HEXACO factor scales, the Altruism facet scale items are not included. However, if factor scores are to be computed as principal components, the Altruism scale can be included along with the other 24 facet scales. (As noted above, this facet generally divides its loadings between Honesty-Humility, Emotionality, and Agreeableness.)

The 60-item HEXACO-PI-R (HEXACO-60)

More recently, we have developed a shorter version of the HEXACO-PI-R in response to demand for an instrument that would be suitable when the time available for personality assessment is very short (Ashton & Lee, 2009). In constructing the HEXACO-60, we decided that each of the six scales should contain 10 items that collectively cover a wide range of content, with at least two items representing each of the four narrow traits of each scale in the longer HEXACO-PI-R. We aimed to construct an instrument that would show moderately high internal-consistency reliability (to the extent permitted by the brevity and breadth of the scales), low interscale correlations, and a factor structure in which items (or facets) of the same broad scale would show their primary loadings on the same factor of a six-factor solution. After selecting the subset of 60 items, we examined the HEXACO-60 in samples of college students and community adults, and found that the instrument did show the desired properties.

Top of the Page