Testing significance of differences between two chi-squares

This paper presents a way to test the difference between twoX ²'s. The test requires evaluating the difference with respect to theT _m (x) Bessel function. Included is a table of the 5 percent and 1 percent points for the Bessel function with degrees of freedom up to 100.This investigation was supported by National Science Foundation Grant Number GE 9846.     

The Vocational Rehabilitation Counselor on Mental Patient Treatment Teams

This study asked the question: What are the role expectancies of treatment personnel in the treatment community and are they consistent across patients' emotional condition? An 83-item inventory was developed and administered under three different sets of instructions, each characterizing a different patient condition. Treatment staff and administrators were asked to rate items on the basis of the level of expectancy they held for their team vocational rehabilitation counselor. The analyses suggest that treatment staff and counselors do not agree on role expectations and that patient condition is a significant mediating variable.     

Reading and typing

As an extension to an earlier study on typing, an experiment was carried out on skilled touch typists in which they were asked to type texts that differed in word structure and word length. It was predicted from certain assumptions about the reading process that there should be significant effects of structure, length and their interaction. These predictions were confirmed.     

Are thresholds reduced by illusions? An attempt at replication

Three experiments are reported, which are attempts to replicate the finding of Ross and Gregory (1964) that difference thresholds for weights can be lowered by means of the size-weight illusion. Three different procedures were used, the first one (experiment I) being designed to show whether or not changes in a subject's judgement criterion could account for apparent changes in sensitivity. The second method (Experiment II) was a replication of Ross and Gregory's first procedure, in which the standard weight was judged before the comparison. In Experiments I and II a larger illusion was induced than in the original studies, but in Experiment III both the weights and container sizes were practically identical to those used by Ross and Gregory. The procedure was also the same as their most successful procedure (number 3) in which standard and comparison were judged simultaneously. The findings were uniformly negative: there was no evidence of criterion shift when the size-weight illusion was induced nor did we find the lowering of threshold previously reported.