tests (Hough et al. 2001 ). While research clearly shows that cognitive ability validly
predicts overall job performance better than any other single characteristic
(Schmidt 2002 ), some question whether the historical focus on task behaviors as
performance (Austin and Villanova 1992 ) has led to the dubious conclusion
that other person characteristics arenecessarilyweaker predictors of all types of
performance, especially if one considers citizenship behaviors (Motowidlo et al.
1997 ). Using personality, biodata, interviews, and other predictors of organizational
citizenship (Borman et al. 2001 ; Organ and Ryan 1995 ) in conjunction with cognitive
ability, researchers have attempted to reduce adverse impact (e.g. Bobko et al. 1999 ;
De Corte 1999 ; Hattrup et al. 1997 ; Murphy and Shiarella 1997 ). Unfortunately, such
eVorts typically fail to reduce impact by a practically meaningful degree while
retaining criterion-related validity (Hough et al. 2001 ; Sackett et al. 2001 ).
Still, researchers have studied non-task behaviors in their own right for two
diVerent reasons. Regarding extra-role (Van Dyne et al. 1995 ) and citizenship
behaviors (which are distinct; Organ 1997 ; Rotundo and Sackett 2002 ), selecting
job applicants for these types of behaviors can, in theory, lead to greater organiza-
tional eVectiveness that is more consistent over time since citizenship supports the
environment in which core job tasks are performed, by deWnition. For instance,
employees who are always willing to assist each other can reduce disruptions in the
Xow of production. The second reason for selecting applicants to perform non-task
behaviors is that today’s organizations often hold multiple goals (e.g. Oswald et al.
2004 ; Rotundo and Sackett 2002 ). The production of raw goods must be balanced
with other concerns such as demonstrating corporate responsibility, for example.
Counterproductive performance behaviors are also related to organizational inter-
ests apart from production or service delivery because they can incur costly damage
(Bennett and Robinson 2003 ; Kelloway et al. 2002 ). Selecting people who will not
engage in absenteeism, theft, sexual harassment, and violence may be critical to
eVective organizational functioning. Although it is evident that such concerns have
existed for years (e.g. most job applications request statements about applicants’
past criminal records), selection theories and practices are now explicitly linking
these goals to individual employee requirements.
Recent notions of performance have also gained depth and complexity with the
inclusion of a time dimension. Theories about adaptive work behaviors attempt to
explain how people can perform well in new or continually changing contexts (e.g.
Pulakos et al. 2000 ) and why the rank order of individuals might change with
experience (Viswesvaran and Ones 2000 ). The selection of adaptive employees is
also gaining usefulness as organizations abandon formal job structures (Cascio
1995 ). With cross-trained teams (Marks et al. 2002 ), employees may need certain
adaptive KSAs that help them decide when and how to perform back-up behaviors
(Dickinson and McIntyre 1997 ) when routine processes are disrupted.
Viewing performance over time has also led to the development of more general
theories. Ployhart et al. ( 2001 ) provided additional support for Sackett and
306 neal schmitt and brian kim