John F. Walsh

Bootstrapping: a tool for clinical research.

The use of the bootstrap sampling technique is applied to the type of data found in clinical research. Confidence intervals are computed for simulated values by use of SAS. By applying this approach, clinical researchers are free to explore topics that do not meet the requirements of traditional statistical analytic methods.

Separability, Welfare Economics and the Theory of Second Best

This paper reconsiders the Lipsey-Lancaster [4] theory of second best and formulates less restrictive sufficient conditions for the preservation of the usual first-order conditions for the non-deviant commodities. In particular, we derive a more general class of separable objective and constraint functions which yield the result that, even in the presence of a Lipsey-Lancaster constraint for deviant commodities, the usual Paretian optimum conditions for the non-deviants continue to be optimal. Consider the maximization problem: max F(x1, ..., xn)

Developmental trends in the use of logical connectives

Children in Grades 2, 4, 6, and 8 were tested on their ability to comprehend and categorize stimuli involving eight logical connectives: conjunction, disjunction, conditional, biconditional, and their negatives. Three distinct groupings of difficulty were apparent. The easy group of connectives (negative implication, negative disjunction, and conjunction) had the common factor of being verbalized with the connective “and.” The second group, showing a definite developmental trend, consisted of the negative biconditional and the disjunctive, which shared the common factor of being verbalized with the connective “or.” The third group revealed an almost total failure to understand implication and biconditional, verbalized with “if,” as well as difficulty in handling the negative conjunctive which verbalizes as a disjunction of negatives. Results were considered to support Bourne’s hypothesis of an intuitive truth-table approach to multiple rule learning rather than Neisser and Weene’s postulation of a hierarchy of difficulty levels.Squirrel monkeys were trained to avoid an airblast in a two-way shuttle procedure. Following acquisition and after a 10-day rest, one group received S days of relearning, the other group received 2 days of relearning and 3 days of training with CS position reversed. Both groups were then extinguished to the CS in its original position. The reversed group exhibited significantly greater resistance to extinction.

On increasing the supply of nurses: An interstate analysis

During a period in which the health care system is undergoing a great deal of change and is subject to shifts in policy, it is necessary to understand the workings of each of its markets. In this paper the nursing market is analyzed. Among the studies in the literature on this subject is that of Benham [1971], whose work has been extended in this research. Benham used state-aggregated data for 1950 and 1960 to estimate a three equation model of supply and demand for registered nurses (RNs). Equations for RN wage, RN stock, and RN labor force participation rate were estimated. The model presented here is extended to include licensed practical nurses (LPNs) who are an important substitute for registered nurses in most practice environments. In addition, the participation rate equations, which have 0-1 range dependent variables, are estimated using the logit transformation. Because the duties, as opposed to training, of the two groups of nurses overlap greatly, simultaneous treatment of the two markets is important. The reaction of each in response to policy changes will determine outcomes. In regulation of the nursing market, especially as it relates to concerns about ameliorating nurse shortages, there are three policy alternatives generally given consideration. The first is to provide nursing educational subsidies that will increase the output of new nurses. The expectation is that nurse stocks will rise and, for a given participation rate, nurse labor supply will increase.

SEX AND ATTRIBUTIONS ON REACTIONS TOWARD ALLEGED SPOUSAL ABUSE VICTIMS

This study examined the emotional and behavioral reactions of 118 predominantly Euro-American, middle-class university students toward an alleged spousal abuse victim. In the context of a decision-making task, participants encountered another male or female student (a confederate) who was supposedly injured in an accident or a spousal abuse incident. Participants completed the Trait Questionnaire, Emotional Reaction Measure, and Social Distancing Measure. Generally, participants felt less sympathy and more anger and ascribed less positive traits to the abuse victim than to the accident victim. In addition, they distanced themselves socially from the abuse victim. Further analyses showed that female participants displayed more negative trait ascriptions and behavioral reactions to the abuse victims than did male participants.

A Follow-up Study of 300 Ex-Clients of a Drug-Free Narcotic Treatment Program in New York City

This article deals with an attempt at evaluating the success of a drug-free narcotic treatment program in New York City through a follow-up study of 300 ex-clients of the program. Two hundred ex-clients out of the 300 target population were actually interviewed. The number of months the clients had remained in the treatment program was found to be a key factor in the outcome of the program. The length of time in the program was itself highly dependent upon the clients background: the higher the pretreatment pathology the lower the chance of treatment success.

A FACTOR ANALYSIS OF MATHEMATICAL GIFTEDNESS

90 mathematically gifted high school students selected on the basis of aptitude and interest in mathematics were compared with 63 unselected college preparatory high school students in the same general ability range. A principal-component method of factor analysis was used to determine the minimum number of independent dimensions needed to account for the variance in the correlational matrix.

Use of a general analysis of variance program in missing data situations

Procedures are described which enable researchers to easily modify a general N-way analysis of variance program so that it can be used in unequal N cases. Advantages in terms of range of application, storage requirements, and accuracy are presented. FORTRAN instructions illustrating the general approach are given.

The effect of stimulus familiarization procedure on paired-associate verbal learning

Ss were familiarized with stimulus words under either of two conditions: aimed at differentiating stimuli or creating associations to them. When the stimuli were used in a paired-associate task, Ss who attended to differentiating stimulus properties performed better than Ss who formed associations to the stimuli or Ss who had no pretraining. Ss who formed associations did not differ in learning from Ss without pretraining.

An efficient procedure for selecting equal probability samples

FORTRAN code that implements an algorithm for selecting equal probability samples is given. The procedure is brief and decides to include or not include a record as it is called. Consequently, storage requirements are minimal. The procedure gives the exact number of values specified, with each record in the population having an equal probability for inclusion in the sample, including the last.