JavaScript is disabled for your browser. Some features of this site may not work without it.
Integrity, conscientiousness, neuroticism, and ability : relationships and measurement : a dissertation presented in fulfilment of the requirements for the degree of Doctor of Philosophy in Psychology at Massey University, Palmerston North, New Zealand
The purpose of this dissertation was to increase knowledge relevant to psychometrically oriented
workplace selection and classification. Multivariate relationships among integrity, conscientiousness,
neuroticism, and fluid and crystallised ability scales were investigated. Adverse impact and the
capacity to use response time information as criteria of ability scoring were also investigated. These
three foci all had the potential to contribute knowledge capable of increasing the accuracy of the
measurement and interpretation of commonly used psychometric assessments.
Two cross-sectional studies were undertaken. The first study used archival data for extant
assessments of ability, general personality, and integrity. It involved 211 participants having
undertaken assessments as a function of job applications. The second study designed and piloted new
scales of integrity, conscientiousness, neuroticism, and fluid and crystallised ability. It involved 317
participants who completed these scales online as voluntary participants.
The first study found integrity to be related to both conscientiousness and neuroticism, but not
substantially related to ability. Conscientiousness was also negatively related to crystallised ability.
These findings were replicated in the second study. The first study’s neuroticism scale which included
a suspicion/cynicism facet (i.e., subscale) had a negative relationship with ability indices. This finding
was not replicated in the second study. This may have been due to the absence of a neuroticism facet
measuring suspicion/cynicism in the second study.
Those identifying as Maori within the first study were found to score substantially less well
than non-Maori on crystallised ability indices, but not other scales measured. Calculations suggested
any resulting adverse impact could be reduced by combining ability assessments with scales of
integrity, conscientiousness, and neuroticism. These calculations were based in part upon the
assumption that relationships among assessments are likely to account for shared variance in job
performance predictions. No significant differences were found in the second study; although the very
small sample size used (N = 22) encourages caution regarding the robustness of this result.
Findings from the second study also suggested that relative to low-ability respondents, highability
respondents took less time to complete crystallised items and more time to complete fluid
ability items. A small significant relationship was also observed between conscientiousness and the
length of time taken to complete the fluid ability scale.
The studies undertaken had a number of limitations. One limitation shared across these
studies was the very small number of participants identifying as Maori (N46 in Study 1 and N22 in
Study 2). Another common limitation was the inability to generalise findings based upon crosssectional
data drawn from participant groups of convenience rather than individuals selected via
probability sampling.
Despite such limitations the preceding findings have a number of practical implications. One
such implication is that relationships among scales may vary according to whether the level of analysis
undertaken is at the Big Five or facet level and whose version of a scale is examined. On this basis
practitioners should examine items in order to understand scale output, and researchers should
examine relationships at the level of facet or ability subcomponent. Practitioners should also use
personality assessments alongside those of ability if they wish to maximise predictive validity and
reduce adverse impact for those identifying as Maori. Furthermore, the use of response time
information in testing is probably better suited to controlling and checking respondents’ approach to
answering assessments than incorporation in scoring algorithms.
This dissertation makes two novel contributions concerning relationships between response
time and participant characteristics. Firstly, negative relationships between ability indices and
conscientiousness or neuroticism scales appear real. They do not appear to be a consequence of more
conscientious or neurotic respondents taking longer to complete ability scales. Secondly, poor timemanagement
strategies do not explain response time results that are inconsistent with the belief that
higher-ability respondents will complete assessments more quickly than their lower-ability peers.
Differences in the cognitive requirements associated with fluid and crystallised tasks instead appear to
explain why higher-ability respondents take relatively less time to complete crystallised scales, but
relatively more time to complete fluid ability scales.