heating conversions in the 12-month period following the
survey. The prediction of 2,062 such conversions represents only 42 percent of the actual. On this basis, intentions measures grossly underestimate actual behavior. If
predictions were based solely on the "within-one-year
intender" group, the prediction would represent only 29
percent (1,420/4,945) of actual conversions.
CONCLUSIONS AND IMPLICATIONS
Conceptual and empirical work in the area of consumer
behavior suggests that it would be fruitful to explore
linkages between attitudes and behavior and between intentions and behavior within consumption contexts. If, for
example, an electric utility could forecast future electrical
demand based on periodic attitude or intentions measures
obtained from a representative sample of its customer base,
it would be better able to match supply capacity to emerging needs. Many utilities, in North America in particular,
have been severely criticized for their past "straight-
line-extension" approach to load growth forecasting.
Implicitly, utilities have used a "past behavior/future
behavior" model. Might a "present intentions/future behavior" model work better?The present study sheds some light on the validity and
practical usefulness of employing intentions measures to
predict future electric space heating conversion behavior.
Several tentative conclusions appear to be warranted.
|1. ||Ability to respond. Homeowners are capable of
expressing their future space heating fuel choice intentions.
Most respondents answered the two-stage intention question
clearly in that they provided a specific fuel form intention
if they intended to switch.|
|2. ||Building an intention-behavior model. Time-accurate
intentions fulfillment rates fall far short of behavior for
one-year planning horizon intenders; only one-fifth of them
engage in the intended behavior action. Although delay of
intention fulfillment is the norm, some householders advance
their intentions fulfillment. This implies that the forecasting of one-year conversions behavior rates must incorporate the impact of earlier-than-expected actions on the
part of intenders. The present study suggests that about
5 percent of the 1-5 year intenders and about 3 percent of
the over-five-year intenders actually advance their conversion to year 1. Actual space heating conversion behavior
can be predicted by constructing an intentions behavior
model including "factoring up" of conversion intention|