The effect of unequal group size on Tukey's HSD procedure

A limitation of the Tukey HSD procedure for multiple comparison has been the requirement of equal number of observations for each group. Three approximation techniques have been suggested when the group sizes are unequal. Each of these techniques was empirically analyzed to determine its effect on TypeI error. Two of the considered variables, average group size and differences in group size, caused differing actual probabilities of TypeI error. One of the three techniques (Kramer's) consistently provided actual probabilities in close agreement with corresponding nominal probabilities.     

The effects of figurai shape on the perception of area

It was hypothesized that figures of equal area and different shapes are perceived as differing in area and that these apparent differences in area are due to the interaction of certain shape variables. Thirty figures were presented to 45 Ss for paired comparisons. Forty-five sets of 30 Choice scores were calculated. The performance of each S was tested for consistency by the Method of Circular Triads (p>.01). A Coefficient of Agreement (u) indicated that the Ss agreed among themselves (p>.01). Fifteen Ss were tested twice for test-retest reliability (p>.01). The 45 sets of Choice scores were then converted to rank values and averaged over all Ss. The mean rank values of the 30 figures were treated with the two-way analysis of variance by ranks to determine the significance of the differences among them (p>.01). A Coefficient of Concordance (W) was calculated on the 45 sets of rank values to support the results of the analysis of variance (p>.01). Subsequently, the 30 mean rank values were treated as the dependent variable in a multiple correlation (R) with the shape variables as the independent variables (p>.01). The apparent area of the figures was found to vary inversely with perimeter, number of turns, breadth, and external area.     

A stochastic model for free recall

A statistical model for verbal learning is presented and tested against experimental data. The model describes a Markov process with a realizable absorbing state, allowing complete learning on some finite trial as well as imperfect retention prior to this trial.This work was carried out while the author was at Lincoln Laboratory, Massachusetts Institute of Technology.Operated with support from the U. S. Army, Navy, and Air Force.