Page 64 - Psychological Management of Individual Performance

progress, problems, and prospects for future research 43
Among ability measures, there is little concern with the assumption of a positive linear
relation between ability score and performance throughout the range of scores. Higher
levels of cognitive ability, for example, are regarded as more desirable for job perfor-
mance regardless of job demands. In contrast, higher levels of extroversion may be
desirable for jobs involving social interaction and team leadership, but may not be de-
sirable for jobs that require intense concentration on solitary tasks.
Tett et al. (1999) argued that meta-analytic procedures used to assess the validities
of personality–performance relations assume linear relations between personality con-
structs and performance throughout the range of responses, and so may cause a distor-
tion in findings. To address this problem, Tett et al. (1999) proposed an extension of
meta-analytic procedures. Applying the proposed procedures to the data set examined
in their previous meta-analysis, they found that validities for several of the personality–
performance relations to be higher than previously reported. Although the effectiveness
of Tett et al.’s (1999) proposed methods for extending standard meta-analysis is a tech-
nical question that requires further evaluation, it is clear that I/O researchers will need
to pay greater attention to issues of non-linearity and bidirectionality in future research.
Meta-analytic studies of ability–performance and personality–performance relations
build upon advances in basic psychology domains, most notably intelligence and person-
ality. Meta-analyses of ability–performance relations during the past 15 years have relied
on long-standing conceptions of intelligence positing a unitary general factor (Spearman,
1904). This conceptualization enables aggregation of studies using often diverse mea-
sures of intelligence or cognitive ability. As such, observed relationships between ability
and performance are interpreted in terms of the influence of general mental ability on
job performance. However, conceptualizations of intelligence that distinguish between
individual differences in crystallized and fluid intelligence (e.g., Cattell, 1963) suggest
that further attention be paid to how such conceptions of intelligence may affect the
generalizability of conclusions based on meta-analyses assuming a unitary conception
of general intelligence, especially when the prediction of performance moves beyond
entry-level jobs and young adults.
In the personality–performance domain, the FFM provided the first comprehensive
organizing framework useful for meta-analytic investigations. Although the FFM has
proved useful, a number of I/O researchers have raised troubling concerns about the
adequacy of FFM and criterion constructs for the current “world of work”. Non-ability
constructs traditionally considered important determinants of performance, such as mo-
tivation, interests, and values, are not easily fitted within the FFM. As a consequence,
research aimed at investigating new constructs, such as personal initiative (Frese, Fay,
Hilburger, Leng, & Tag, 1997), proactive personality (Seibert, Crant, & Kraimer, 1999),
and motivational traits and skills (Kanfer & Heggestad, 1997) tend to receive less atten-
tion in I/O meta-analytic research employing the FFM taxonomy.
Similarly, the reliance on archival data, in which job performance is often undiffer-
entiated with respect to the aspects of the criterion domain tapped by the performance
measure, is also problematic. Recent advances in the criterion domain, distinguishing
between job performance with respect to technical (Campbell et al., 1995), contextual
(Motowidlo & Van Scotter, 1994), and adaptive (Pulakos, Arad, Donovan, & Plamandon,
2000) dimensions have not been adequately addressed in meta-analytic research. For
research to evaluate the role of person constructs not readily captured within the FFM