26.4 THREE PRELIMINARY ANALYSES
(a) The Data. Quantitative records of objectives kept over time and revised after re-
ceiving reports of performance with respect to those objectives have been collected
at several sites. Regressions to estimate the parameters in equation (2) are reported
below. What follows is a brief outline of several research questions related to the
measurement of performance at this company along with a presentation of results
from some preliminary analyses. In discussing these three questions, data reported
will concern six different measures used at four different sites.
(b) Empirical Confirmation of the Adaptive Process. A first obvious question is
whether the attainment discrepancy model, confirmed using experimental subjects^37
and groups of executives and master’s degree students taking a graduate marketing
course,^38 will be confirmed using field data. To test this, data from the goals of the
retail units for total deposits at one site were collected. At this particular site, the dis-
tribution of deposit goals, which were stated in local currency, seemed very skewed;
examination of the distribution of the log of these goals revealed a distribution that
better conformed with the assumptions of the general linear model. In addition, sev-
eral of the retail units had recently been converted from being franchise operations to
being wholly owned units of the company. This change in status was indicated by a
dummy variable called Branch Conversion in the analysis. The results are reported in
Exhibit 26.1; as can be seen by inspection of the results, they are roughly consistent
with those found in the past research. First, there is a significant, positive constant;
0 is greater than zero at the 0.01 level. Second, there is a significant effect from pre-
vious aspiration level; as predicted, 1 is greater than zero, also at the 0.01 level.
Third, performance in the previous period has a significant effect; 2 is greater than
zero at the 0.05 level. Finally, the control variable for the conversion of retail units is
also significant; the parameter measuring the effect of being among those observa-
tions 3 indicates that they have a slighltly higher intercept than the other observa-
tion (p < 0.05).
26 • 8 DYNAMIC PERFORMANCE MEASUREMENT SYSTEMS FOR A GLOBAL WORLD(^37) Glynn, Lant, and Mezias, 1991.
(^38) Lant and Montgomery, 1987; Lant, 1992.
Measure:Annual branch goals for liabilities stated in local currency.
Data:102 usable observations were obtained.
Predictor Variable Coefficient
Constant 2.270
Log of Previous Aspiration Level .452
Log of Previous Performance Level .078
Branch Conversion .177
*p < 0.05 **p < 0.01
Restated Model:
LogGoalt= 2.270 + .530 LogGoalt–1+ .078 (LogPerft–1– LogGoalt–1) + .177 BrConv
Exhibit 26.1. Goals for Deposits.