contrast with boundary spanning units is extreme. Thompson described the contrast
as follows: “Whereas coordination is a central problem for the technical core of the
organization, adjustment to constraints and contingencies not controlled by the or-
ganization is the crucial problem for boundary spanning components.”^42
In terms of our study, the units that provide service and support to the retail units
are closer to the technical core. There is no entry and exit of customers as exists in
the retail unit; contact with customers is channeled in ways that categorize the prob-
lem and route the contact to the point most equipped to handle the inquiry. Thus, we
would expect that the goals of the units in the technical core would have some of the
same characteristics that Lant^43 ascribed to aspirations under a rational expectations
process. Based on her discussion, this would imply that the coefficient on previous
aspiration level would be zero. By contrast, the retail units are directly accessible to
nonemployee customers whose inquiries cannot be effectively controlled or routed in
person as they can by remote communication. Thus, we would expect these units to
follow a process as specified in equation (2).
To examine aspirations at the technical core, we examined the volume of manned
calls per month at one of the sites where the unit under study engaged in solely a sup-
port function; the results are reported in Exhibit 26.5. To examine the boundary span-
ning units, we once again examined the sales volume of a consumer product at a site
that had retail operations. Control variables were also included for size (measured in
units of 10 millions of local currency) and branch manager tenure in the organization
(measured in quarters). The results are reported in Exhibit 26.6. The negative con-
stant is not significantly different from zero, and there is a significant, positive coef-
ficient for both previous aspiration level and previous performance level, a signifi-
cant positive effect for branch size, and a nonsignificant effect for tenure. Taken
together, Exhibits 26.5 and 26.6 show that, as predicted, the coefficient on previous
aspiration is not significant for the unit at the technical core, while it is significant for
goals set by units that are at the boundary of the organization.
26.5 CONCLUSION. As indicated, the results here are only preliminary. We are not
yet ready to present results with sufficient precision to justify a thorough review of
the larger question of how performance evaluations are changing in response to in-
26.5 CONCLUSION 26 • 11Measure:Monthly goals for divisional performance on number of manned calls.
Data:27 usable observations were obtained.
Predictor Variable Coefficient
Constant 7.764*
Log of Previous Aspiration Level .028
Log of Previous Performance Level –.307
p < 0.05 **p < 0.01
Restated Model:
LogGoalt= 7.764 + .279 LogGoalt–1+ .307 (LogPerft–1– LogGoalt–1)
Exhibit 26.5. Goals for Manned Calls (technical core).
(^42) Id., p. 81.
(^43) Lant, 1992, pp. 627–628.