Nonorthogonal two-way analysis of variance.

Examined the advantage of disadvantages of 5 different exact analyses of variance for nonorthogonal 2-way designs with respect to orthogonality of the analyses, parametric hypotheses tested, and model comparisons made by the analyses. It is proposed that experimenters, when faced with the necessity of performing a 2-way ANOVA, carefully consider these analyses with regard to the a priori information they have about the data, the questions they expect the analysis to help answer, and the questions each analysis is best equipped to answer. It is also suggested that experimenters choose the analysis that best fits their needs rather than depend on one for all situations. (16 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Sample size and optimal designs for reliability studies

Insulin-like growth factors (IGFs) I and II stimulate growth and expression of specific genes through binding to cell membrane receptors. IGF binding proteins also bind IGF-I with higher affinity than the receptor. They are found in the circulation and tissues and can modulate IGF actions. Human IGFBP-1 is phosphorylated on serine residues, which increases its affinity for IGF-I. An acidic, presumably phosphorylated, form of human IGFBP-1 inhibits IGF-I-stimulated DNA synthesis in cultured cells, while a less acidic, unphosphorylated form potentiates this function. Phosphorylation of human IGFBP-3, however, does not affect its affinity for IGF-I. Previously we found that multiple forms of rat IGFBP-1 are obtained by anion-exchange chromatography, raising the possibility that it also is phosphorylated, which led us to examine its properties. Phosphopeptide analysis of 32P-labeled, immunoprecipitated rat IGFBP-1 synthesized by H-4-II-EC3 rat hepatoma cells indicated that it is phosphorylated on two sites that were deduced to be ser107 and ser132 in the central nonconserved domain. Dephosphorylation of purified phosphorylated rat IGFBP-1 did not affect its affinity for IGF-I or its specific binding activity, and the dephosphorylated form inhibited DNA synthesis in 3T3 cells. Incubation of cells labeled with radioactive proline in the presence of monensin and brefeldin A, which inhibit secretion at different sites, led to intracellular accumulation of the least phosphorylated form of rat IGFBP-1, but prevented further phosphorylation. The results suggested that phosphorylation occurs at two sites in cells, the cis-Golgi and the trans-Golgi network. In summary, these studies have shown that rat IGFBP-1 is phosphorylated on two sites by reactions that occur in different secretory organelles and that similar to human IGFBP-3, but unlike human IGFBP-1, phosphorylation does not affect its affinity for IGF-I.     

Balanced designs in longitudinal population pharmacokinetic studies

Despite the proliferation of research on primary nursing, most studies have investigated the effects of primary nursing on quality of care, job satisfaction and collegial relationships. Few researchers have attempted to isolate the key dimensions of primary nursing in an observable and measurable form. The purpose of this exploratory study was to devise an audit tool for primary nursing that builds on Mead's research and incorporates views of patients/clients, relatives, nurses and other members of the multidisciplinary team. Preliminary findings suggest that the audit is a useful indicator of the extent to which the crucial elements of primary nursing are present or absent in any clinical setting. Furthermore it suggests that while many wards claim to practise primary nursing there is a considerable variation in the extent to which this is actually carried out.     

Single-case research designs: A neglected alternative.

Comments on the article by L. Wilkinson & the Task Force on Statistical Inference (see record 1999-03403-008) which was meant to initiate discussion in the field about changes in current practices of data analysis and reporting. The report was concerned with the use of statistical methods only and is not meant as an assessment of research methods in general. The present article criticizes Wilkinson's article, noting that it lacks any recognition that single-case research design might be an alternative to current practices. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Transfer designs and formulas.

This paper includes a comprehensive analysis of the several experimental designs available for the study of transfer of training. The discussion is restricted to the experimental designs and formulae that are appropriate in 2 types of transfer studies—verbal transfer, and predifferentiation, not including motor transfer. The uses and misuses of the various designs and transfer formulae are presented. 78 references. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Transformations for within-subject designs: A Monte Carlo investigation.

Explored the use of transformations to improve power in within-S designs in which multiple observations are collected for each S in each condition, such as reaction time (RT) and psychophysiological experiments. Often, the multiple measures within a treatment are simply averaged to yield a single number, but other transformations have been proposed. Monte Carlo simulations were used to investigate the influence of those transformations on the probabilities of Type I and Type II errors. With normally distributed data, Z and range correction transformations led to substantial increases in power over simple averages. With highly skewed distributions, the optimal transformation depended on several variables, but Z and range correction performed well across conditions. Correction for outliers was useful in increasing power, and trimming was more effective than eliminating all points beyond a criterion. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Modification of smoking behavior: Good designs ineffective treatments.

Cigarette smoking provides opportunities to investigate behavior change processes in a meaningful, naturalistic context that still permits adequate measurement and controls. Although several recent behavior modification studies, exemplified by the A. R. Marston and R. M. McFall (see PA, Vol. 46:Issue 2) paper, are well designed, the treatments employed remain ineffective. It is suggested that treatment ineffectiveness may be due to insufficient pilot (clinical) work and premature freezing of techniques into group programs. It is also argued that process studies of smoking modification must await demonstration of an outcome effect or, at the least, process variables must be shown to effect outcome. The problems involved in verifying smoking rates and whether Ss are performing the "homework" assignments frequently given by behaviorists are also discussed. (25 ref.) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Modification of smoking behavior: Good designs inaccurate reporting.

Discusses inaccuracies in A. Marston and R. Mcfall's (see pa, vol. 46:3232) replication of J. Resnick's (see pa, vol. 43:876) study of smoking behavior. The importance of accurately reporting aspects of methodology in a replication and extension study is noted. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Concurrent and predictive validity designs: A critical reanalysis.

Historical and contemporary discussions of test validation cite 4 major criticisms of concurrent validity that are assumed to seriously distort a concurrent validity coefficient. These include "missing persons," restriction of range, motivational and demographic differences between present employees and job applicants, and confounding by job experience. As a result, predictive validity has been espoused as a scientifically superior strategy for the validation of personnel selection tests. However, it is argued here that frequently the conceptual distinction between predictive and concurrent validity has been exaggerated. More importantly, the differences that may exist have never been shown to render concurrent validity inaccurate as an estimate of predictive validity. Existing data suggest that these differences, if present, have a minimal impact on the magnitude of an obtained validity coefficient. (29 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Repeated measurements designs and counterbalancing.

"Six types of analysis of repeated measurements designs are indicated. The effects of order, interactions containing order, and correlated observations on the components of variance and analysis of variance tests of significance are considered. The first two act, in general, to inflate the error estimates and thus to increase the probability of a Type II error. The correlated observations (if unequal) have the opposite effect, i.e., increase the probability of a Type I error." From Psyc Abstracts 36:01:1AF46G. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Proportional effects in experimental designs.

If the same stimulus set constitutes the levels of both the row and column factors of a 2-factor stimulus design, then, under certain conditions, the row and column effects for design will be proportional. A maximum-likelihood test of this proportionality is provided. Within the context of functional measurement, the maximum-likelihood estimates of the proportional effects provide an efficient interval scaling of the stimulus set. (16 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Analysis of nonorthogonal fixed-effects designs.

Discusses nonorthogonal fixed-effects experimental designs using both the full-rank and reduction in error sums of squares conceptualizations of data analysis. The hypotheses tested by several commonly used methods of analysis are clarified, and suggestions for choice of the most appropriate procedure are proposed. (17 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Analysis of variance for repeated measures designs: Two approaches.

Discusses the relationship between 2 basic models for analysis of variance for repeated measures: the "traditional" model, which involves the assumptions of equal correlations between all variables both within and between S groups, and a less restrictive model based on a multivariate analysis of variance technique which is considered as an alternative to the traditional model. Suggestions are made for efficient computer programing for the traditional model. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Reliability and Sample Size.

Asserts theat a recent comment by Wallin (Amer. Psychologist, 1954, 9, 203) suggests that editors might relieve the overload of publication material by refusing experimental articles in which conclusions are based upon a small number of cases. While it is true, as in the three cases cited by Wallin, that some experimental articles either state or imply generalizations far beyond the limits of their sampling techniques, it appears to the writer that sample size, per se, is not the critical factor. To make sample size a criterion for rejection would be, in effect, to force a return to older, large sample techniques which rnethodologists and statisticians have done so much to streamline in recent years. Such methods, while they offer the comfort of security in numbers, are remarkably inefficient and, in most cases, the additional labor adds no information which could not have been gained through the proper handling of observations of a smaller sample. It should be made clear that a difference, statistically reliable at the .01 level of confidence and based upon a sample of ten observations, is just as stable as a difference at the same level of confidence based upon a sample of 1,000 observations, provided that conditions of randomization obtain. It seems to the writer that the sampling technique employed is more important than the sample size. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Across the Secretary's Desk: Sample of Outgoing Mail.

In February, the Executive Secretary hit upon a new and to him a nicely adaptive mechanism for meeting the constantly felt need to let American Psychological Association (APA) members know at least as much as they want to about the operation of their Central Office. He asked the professional staff of the Central Office to look through copies of the last month's correspondence from the Secretary's Audograph to see if they could find material that was (a) of at least mild potential interest to a number of APA members, (b) illustrative of some facet of APA functioning, (c) free of any threat to the privacy of the recipient, and (d) at least moderately literate. Each member of the staff found some material which he was willing, under pressure, to mark as meeting the above criteria. Though the level of reliability was nothing to brag about, the process, abetted by some editorial license, resulted in the following array of sentences and paragraphs from the Executive Secretary's outgoing correspondence for a month or so early in 1955. Each item is reproduced in this article with the permission of its recipient. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Cultural influences on preference for designs.

An attempt is made to determine the extent to which the frequently demonstrated agreement between individual aesthetic choices is really a general factor related to the aesthetic material and independent of social and cultural influences. "Eight West African designs were shown to 56 West Africans and 56 English Ss each of whom was asked to say which two he liked best and which two he liked least. Within each cultural group considerable agreement was shown by the subjects; since, however, the preferences of the two groups were quite distinct, little evidence was found of a general agreement which depended on the designs rather than on the judges." (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Exact probability treatments of factorial designs.

An exact probability analysis of data which represent frequency of use in various response categories is described. The test can be generalized for any number of classifications and to any order interaction. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Pretesting effects in retrospective pretest-posttest designs.

Two studies evaluated communication skills training by using a pretest-posttest design, including retrospective pretest ratings, to control for response shift bias. A response shift is a change in a subject's internal standard for determining his or her level of functioning on a given dimension. In Exp 1, Ss were 37 hospital employees. Data indicated that the self-report pretest exerted a clear effect on subsequent self-report posttest and retrospective pretest ratings. Training was ineffective and a response shift did not occur. Experimental Ss could not remember and control Ss could remember their pretreatment ratings to a reasonable extent. In Exp 2, Ss were 58 3rd-year dental students. Results show that the training was effective. A behavioral pretest administered prior to the self-report pretest prevented a response shift from occurring. This finding gives empirical support to the contention that Ss' lack of sufficient information about their level of functioning at pretest may be a causal determinant of the response shift. Data also indicate that the retrospective pretest is robust for procedural differences in administering this instrument. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Psychic determinism versus freedom of choice: Order versus chaos.

In response to L. E. Tyler's (see record 1974-25495-001) article on a hopeful psychology, S. Morganstein proposed that orderly thought or behavior, based on a sequence of choices, is either determined by the external environment or a combination of the latter plus internal causes or control. The author submits that a psychologist who recognizes his/her own motivations is freer than one who judges by what society deems profitable. (PsycINFO Database Record (c) 2010 APA, all rights reserved)