Four Perspectives on . . .
18
in the first two rows reveals coefficients that are very similar to one another. When acceleration
is entered first, it is statistically significant, but velocity is not (Equation 1). When velocity is
entered first, it is statistically significant, but acceleration is not (Equation 2). Although velocity
is a slightly better predictor than acceleration (a difference of .001), the analysis was unable to
provide clear evidence in favor of one perspective over the other due to the degree of collinearity
between the two predictors (r = .87).
...TABLE 3 ABOUT HERE...
In the center of the table, velocity is contrasted with deceleration. In Equation 1, velocity
is significant when it is entered first, and deceleration, when entered second, is not. In Equation
2, deceleration is significant when it is entered first, but velocity contributes significant
predictive power even when it is entered in the second block. Thus, these results favor velocity
over deceleration.
The results of the final comparison (acceleration vs. deceleration) are given in the lower
third of Table 3. As would be expected on the basis of the previous two analyses, acceleration is
the better predictor. It is significant in Equation 1, whereas deceleration is not, and it contributes
additional predictive power above and beyond deceleration in Equation 2.
In sum, these results suggested mixed support for H4 (i.e., acceleration) and H5 (i.e.,
velocity). Both variables predicted persuasion, but it was impossible to discriminate between
them in these data. Despite a significant bivariate relationship between deceleration and
persuasion (see Table 1), we rejected H6 on the grounds that deceleration contributed no unique
information to the regression models in any of the comparisons.
3
Discussion
Convention