Unequal intervals and unequal n in trend analyses.

Abstract

The method of deriving orthogonal coefficients for the general case of unequal n and unequal intervals between points of the quantitative independent variable is demonstrated.

Adaptation-level theory: Absolute or relative anchor effectiveness?

“…As expected from previous results (see above), both parametric (Gaito, 1965) and nonparametric (Ferguson, 1965) trend analyses demonstrated that the quadratic and bitonic components of the anchor-AL relationships were statistically significant, both for the S00- again, is inconsistent with Helson's model (1947Helson's model ( , 1964) is graphically represented in Fig. I (heavy series) and Fig.…”

Decision time, subjective probability, and task difficulty

“…Clearly, the type of information processed by a subject varied little from decision to deci- appear that the presence of nonlinearity in the data may largely be attributable to the response times for 50% judgments, which were discrepantly low. A post hoc trend analysis (Gaito, 1965;Kirk, 1982) (Scheffe, 1953) showed that the 50% response times were not significantly shorter than those for the next highest subjective probability category, 51 %-60% [F(1,240) These results were inconsistent with Phillips and Wright's (1977) sequential model of probabilistic processing, in that the decision time of a 50% response was no shorter than that of other probabilistic responses, excluding 100% responses. There was also only moderate support for a strict monotonic decreasing relationship between binary decision time and associated confidence or probability over the half (50%-100%) range of probability.…”

Width of the latitude of acceptance as a determinant of attitude change.

“…Change differed significantly from zero only in the wide latitude condition (t = 2.16, df = 20, p < .05). Neither the main effect of discrepancy size (F = 1.84) nor the interaction between latitude width and discrepancy size (F = 1.10) was significant, and the quadratic trend (Gaito, 1965) observed among the wide latitude subjects only approached significance (F=2.88, df= 1/109, £<.10). e To determine whether the attitude change findings might be test-retest phenomena unrelated to the reception of a message, attitude change on each of the irrelevant issues (gun control and legalization of marijuana) was examined as a function of latitude of acceptance width on the birth control issue, but no signficant relationships were obtained.…”