In this paper, we set out to examine the relationships between expected consequences of growth (and some control variables) on the one hand, and over-all growth willingness on the other. In doing that, we used data from three separate survey studies that all had employed the same measuring instruments. This allowed us to come up with many useful comparisons. Some results recur very consistently across sub-analyses and can therefore be accepted even if the associated probability of the coefficient is not very low in every analysis. Other results, while statistically significant in one analysis, may be disregarded because the result appears in isolation. The general approach--to repeat the same measurement in several separate surveys--is something we would highly recommend to other researchers. Results from multiple samples are a much better basis for determining ones degree of confidence, than is significance testing alone.
Our regression models could explain about 25 percent of
the variation in growth willingness in our different sub-groups.
This shows that there are substantial relationships between
expected consequences of growth and growth willingness.
Nonetheless, three quarters of the variance is left unexplained
by our models, and the reasons for this deserve some discussion.
A first possibility is, of course, that important explanatory
variables were omitted. Perhaps other expected consequences of
growth are important and therefore should have been included. As
our selection of explanatory variables was based on a thorough
literature review as well as pilot interviews we regard it highly
unlikely, however, that any new variables that are more important
than those currently included are to be found. The control
variables we have used are also the most important ones in
previous research on small firm growth. Therefore, the add
variables route is unlikely to raise explained variance by
more than a few percentage points.
© 1997 Babson College All Rights Reserved
Last Updated 06/01/98