Content area
Full Text
Subordinate and supervisor leader-member exchange (LMX) were examined as correlates of delegation and as moderators of relationships between delegation and subordinate performance and satisfaction. Raw score analyses of data on 106 dyads showed both to be significantly related to delegation and to have similar main and moderating effects for subordinate performance and satisfaction. Finally, within- and betweengroups analyses largely supported the level-of-analysis predictions of the LMX approach. Implications for future LMX research are discussed.
As Graen and Uhl-Bien noted, "Research into Leader-Member Exchange (LMX) theory has been gaining momentum in recent years, with a multitude of studies investigating many aspects of LMX in organizations" (1995: 219). Originally beginning as vertical dyad linkage (VDL) theory (Dansereau, Graen, & Haga, 1975), this approach challenged the prevailing view of the time, the average leadership style (ALS) model, which treated leader behavior as being reasonably consistent toward all subordinates.
Theoretically, the LMX-VDL approach specified that "the appropriate level of analysis is not the work group. . . but the vertical dyad" (Graen & Cashman, 1975: 150), and operationally it required that within-groups relationships, whose presence supports the LMX approach, be stronger than between-groups relationships, which support the ALS approach (Graen, Liden, & Hoel, 1982: 868). Further discussion of and elaboration on LMX and expected within- and between-groups comparisons can be found in, among others, Dansereau (1995: 483-484), Dansereau, Alutto, Markham, and Dumas (1982: 85), Ferris (1985: 777), Katerberg and Hom (1981: 218-219), Nachman, Dansereau, and Naughton (1983: 171-172), Vecchio (1982: 200201, 206), Williams, Podsakoff, and Huber (1987: 3-6), and Yammarino, Dansereau, Markham, and Alutto (1980: 193).
Several dozen empirical investigations of the basic LMX model have been published since 1972 (Dienesch & Liden, 1986; Graen & Uhl-Bien, 1995) and, conceptualized as described above, leadermember exchange has been defined as "(a) a system of components and their relationships (b) involving both members of a dyad (c) in interdependent patterns of behavior (d) sharing mutual outcome instrumentalities and (e) producing conceptions of environments, cause maps, and values" (Scandura, Graen, & Novak, 1986: 580). Support for this approach has come from field studies that suggest that leader-member exchange may predict outcomes such as the severity of job problems (e.g., Dansereau et al., 1975), job satisfaction and performance (e.g., Graen, Novak, &...