1 OF 118

AU Dreger-Ralph-Mason.

TI Sorting data sets and computing medians for skewed distributions.

SO Educational & Psychological Measurement. 1995 Oct Vol 55(5) 785-790.


ID automation of computation of medians for ordered & unordered sets of scores & for skewed distributions with program called SHELMEDS.BAS.

AB Indicates that clusters derived from data sets require meaningful interpretation, for which standard statistical packages, major spreadsheets, numerical computation texts, and the specialized Journal of Classification provide little help. Interpretation can be made from a profile of summary scores, means, or medians, representing distributions of scores. For such a profile, medians are preferable to means where the distributions are highly skewed. To automate computation of medians in either ordered or unordered sets of scores, a program, SHELMEDS.BAS, was written incorporating the very fast Shell-Metzner sort algorithm and routines for odd or even numbers of scores. (PsycINFO Database Copyright 1996 American Psychological Assn, all rights reserved). ************************************************************************

2 OF 118

AU Ogasawara-Haruhiko.

TI The gamma-gamma regression for the distribution model of event times.

SO Japanese Psychological Research. 1995 Vol 37(2) 70-79.



ID gamma-gamma regression model for distribution of event times, application to attention testing.

AB Proposes the gamma-gamma regression model (GGRM) of event times, in which continuous covariates together with the random (individual difference) parameter lambda are involved, corresponding to the extensions of the Poisson model. The derivation of the GGRM from the gamma regression model is provided, and the method of estimating parameters is described. Marginal maximum likelihood estimates for the hyperparameters in the distribution are given by numerical methods. The model is applied to 2 data sets involving attention tests and is compared with the results of a lognormal regression model. Results indicate that the GGRM is comparable to the lognormal regression model. (PsycINFO Database Copyright 1996 American Psychological Assn, all rights reserved). ************************************************************************

3 OF 118

AU Wedell-Douglas-H.

TI Contrast effects in paired comparisons: Evidence for both stimulus-based and response-based processes.

SO Journal of Experimental Psychology: Human Perception & Performance. 1995 Oct Vol 21(5) 1158-1173.



ID context independent scale values vs contextual parameters tied to features of stimulus distribution vs responses associated with paired distributions, paired contrast probabilities, college students.

AB Participants made paired comparisons of square sizes, with blue squares presented on the left side of a computer screen and red squares on the right. Context was manipulated by varying the distributions of blue and red squares separately. In Exp 1, target squares from low-range distributions were judged larger than the same-size squares from high-range distributions. In Exp 2, ranges were equated but ranks manipulated between distributions, pairing bell with U -shaped and positively skewed with negatively skewed distributions. Results provided little support for a rank-dependent valuation model. Exp 3 used distributions that were designed to test between adaptation-level and response-equalization models. Both models received support, with response latencies constituting an important moderating variable. Response patterns for short latency participants were consistent with a stimulus-based adaptation-level process, and those for longer latency participants were consistent with a response-equalization process. (PsycINFO Database Copyright 1996 American Psychological Assn, all rights reserved). ************************************************************************

4 OF 118

AU Hopkins-William-G. Green-Jack-R.

TI Combining event scores to estimate the ability of competitors.

SO Medicine & Science in Sports & Exercise. 1995 Apr Vol 27(4) 592-598.


ID validity of sports ability measures, simulation assessment of competitive potential.

AB Investigated validities of 9 measures of ability derived from scores of 2 or more competitive events. Simulations were performed for different numbers of competitors, events, and event entries for a range of validity of performance in a single event. A complete set of simulations was repeated for normal distribution of competitors' ability; skewed distribution of ability; event validity related to ability; validity, ability, and spread of scores differing between events; and events differing in difficulty. The mean raw scores were generally the most valid measure. The personal best was comparable to the mean only when the number of entries approached 1 per competitor. The least-squares mean of raw scores had highest validity when events differed in difficulty and should be used when events differ in length, or when event scores are affected by environmental conditions, judging bias, or uneven matching of competitors in match-play sports. (PsycINFO Database Copyright 1995 American Psychological Assn, all rights reserved). ************************************************************************

5 OF 118

AU Egan-Michael-F. Hyde-Thomas-M. Kleinman-Joel-E. Wyatt-Richard-Jed.

TI Neuroleptic-induced vacuous chewing movements in rodents: Incidence and effects of long-term increases in haloperidol dose.

SO Psychopharmacology. 1995 Jan Vol 117(1) 74-81.



ID stable vs increasing dosage of haloperidol vs withdrawal period, vacuous chewing movements, male rats, animal model of tardive dyskinesia.

AB Used the vacuous chewing movement (VCM) syndrome as a model of tardive dyskinesia (TD) to examine the ability of increased neuroleptic doses to produce long-term suppression of dyskinetic movements in 53 rats treated for 15 wks with haloperidol decanoate. For the next 21 wks, half the group received a 50-150% increase in dose while the other half received the same dose. Ss were also followed during a 28-wk withdrawal period. Total VCM ratings showed a skewed distribution, with some rats exhibiting few movements and others developing marked and persistent movements. Increasing doses did not suppress VCMs, nor did they exacerbate movements during the withdrawal period. To the extent that the VCM syndrome models TD, the absence of long-term suppression of the VCM syndrome suggests that increasing depot neuroleptic doses may not be a useful long-term strategy for TD suppression. (PsycINFO Database Copyright 1995 American Psychological Assn, all rights reserved). ************************************************************************

6 OF 118

AU Zimmerman-Donald-W.

TI Simplified interaction tests for non-normal data in psychological research.

SO British Journal of Mathematical & Statistical Psychology. 1994 Nov Vol 47(2) 327-335.


ID main effects & interaction of modular ranks vs detection & downweighting of outliers vs rank transformation procedure for nonnormal distribution data.

AB Studied modular ranks (MRs) and detection and downweighting of outliers (DDOs) and compared them with rank transformation (RTM). Three combinations of treatment effects (main effect, interaction, and both main effects and interaction) were simulated in 2 * 2 factorial designs. ANOVA was performed on the initial scores, and the F ratios representing main effects and interaction in factorial designs were evaluated. The distribution shapes included normal, mixed-normal, exponential, Cauchy, Laplace, uniform, and mixed-uniform distributions. The RTM produced a robust and powerful significance test for interaction for all 7 distributions. Conversion to MRs was just as effective as the transformation to ordinary ranks and had one-fourth the number of values. The DDOs procedure was frequently superior to the rank methods. Implications for the role of scales measurement and nonparametric methods in psychological research are discussed. (PsycINFO Database Copyright 1995 American Psychological Assn, all rights reserved). ************************************************************************

7 OF 118

AU Wilcox-Rand-R.

TI Computing confidence intervals for the slope of the biweight midregression and Winsorized regression lines.

SO British Journal of Mathematical & Statistical Psychology. 1994 Nov Vol 47(2) 355-372.


ID bootstrap vs inferential methods for computing confidence intervals for slopes of biweight midregression vs Windsorized regression lines.

AB Investigated a method of computing confidence intervals for the slope of 2 resistant regression methods, the highly resistant biweight midregression (BMR) and Winsorized regression (WR). Two bootstrap methods, resampling the residuals and resampling the pairs of points for computing confidence intervals, were found to be highly unsatisfactory. An inferential method performed reasonably well when using BMR, and fairly well for WR except for highly skewed distributions with very heavy tails. There were no instances where WR was found to have substantially less power than ordinary least squares (OLS), while in some cases WR had substantially more power. There were situations where the BMR had substantially more power than OLS, but when the predictor had a highly skewed distribution, there were situations where the reverse was also true. (PsycINFO Database Copyright 1995 American Psychological Assn, all rights reserved). ************************************************************************

8 OF 118

AU Algina-James. Oshima-T-C. Lin-Wen-Ying.

TI Type I error rates for Welch's test and James's second-order test under nonnormality and inequality of variance when there are two groups.

SO Journal of Educational & Behavioral Statistics. 1994 Fal Vol 19(3) 275-291.


ID Type I error rates for t vs approximate degrees of freedom vs 2nd order tests, 2 independent samples.

AB Type I error rates were estimated for 3 tests that compare means by using data from 2 independent samples: the independent samples t test, B. L. Welch's (1938) approximate degrees of freedom test, and G. S. James's (1954) 2nd-order test. Type I error rates were estimated for skewed distributions, equal and unequal variances, equal and unequal sample sizes, and a range of total sample sizes. Welch's test and James's test have similar Type I error rates and tend to control the Type I error rate as well or better than does the independent samples t test. Results provide guidance about the total sample sizes required for controlling Type I error rates. (PsycINFO Database Copyright 1995 American Psychological Assn, all rights reserved). ************************************************************************

9 OF 118

AU Bradley-Drake-R. Fleisher-Courtney-L.

TI Generating multivariate data from nonnormal distributions: Mihal and Barrett revisited.

SO Behavior Research Methods, Instruments & Computers. 1994 May Vol 26(2) 156-166.


ID algorithm for generation of multivariate data from nonnormal distribution.

AB An algorithm described by F. A. Graybill (1969) factors a population correlation matrix, R, into an upper and lower triangular matrix, T and T', such that R = T'T. The matrix T is used to generate multivariate data sets from a multinormal distribution. When this algorithm is used to generate data for nonnormal distributions, however, the sample correlations are systematically biased downward. The authors describe an iterative technique (IT) that removes this bias by adjusting the initial correlation matrix, R, factored by the Graybill algorithm. The method is illustrated by simulating a multivariate study by W. L. Mihal and G. V. Barrett (1976). Large- N simulations indicate that the IT works: multivariate data sets generated with this approach model both the univariate distributions of the individual variables and their multivariate structure (as assessed by intercorrelation and regression analyses). (PsycINFO Database Copyright 1994 American Psychological Assn, all rights reserved). ************************************************************************

10 OF 118

AU Schene-Aart-H. van-Wijngaarden-Bob. van-Baaren-Marcelis-J-W.

TI Randomization and generalization in psychiatric research.

SO International Journal of Methods in Psychiatric Research. 1994 Apr Vol 4(1) 1-5.



ID sample randomization, generalizability of relative effectiveness of partial vs full time psychiatric hospitalization, psychiatric patients, Netherlands.

AB Conducted a generalized, longitudinal study of 222 patients comparing the relative effectiveness of partial and full-time psychiatric hospitalization to determine whether generalizability in naturalistic psychiatric research would be maximized if the experimental sample was specific in detail and randomization-related bias (including nonresponse, misuse, and attrition) was minimized. Although randomization was monitored throughout the investigation, a marginally skewed distribution of referring parties was discovered. The internal validity did not deteriorate, but the external validity may have been affected. Researchers should be aware of and alert to the various problems that can threaten the validity of their results. (PsycINFO Database Copyright 1994 American Psychological Assn, all rights reserved). ************************************************************************

11 OF 118

AU Ramsey-Philip-H.

TI Testing variances in psychological and educational research.

SO Journal of Educational Statistics. 1994 Spr Vol 19(1) 23-42.


ID R. G. O'Brien's vs M. B. Brown & A. B. Forsythe's procedure for testing variances with kurtic distributions, robustness & optimal power of simulated data.

AB A review of the literature shows that robust procedures for testing variances have become available. The 2 best procedures have been proposed by R. G. O'Brien (see PA, Vol 66:271) and by M. B. Brown and A. B. Forsythe (1974). An examination of these procedures for 9 simulated populations confirms their robustness and indicates that optimal power can usually be obtained by applying a test for kurtosis. (PsycINFO Database Copyright 1994 American Psychological Assn, all rights reserved). ************************************************************************

12 OF 118

AU Cordes-Richard-E.

TI The effects of running fewer subjects on time-on-task measures.

SO International Journal of Human-Computer Interaction. 1993 Oct-Dec Vol 5(4) 393-403.



ID reduced sample size, accuracy & stability of mean vs median in time on task measures in simulated usability evaluations with skewed distribution.

AB Examines the impact of using fewer Ss on the accuracy and stability of the mean and median in time-on-task measures. Monte Carlo simulations were conducted to model the skewed task completion times typically found in usability studies. Results show that the mean was a more accurate estimate of its respective population parameter than the median and should be the preferred metric in usability evaluations. However, the mean cannot always be used because it will underestimate the population mean when there are missing data as a result of Ss exceeding time limits. Alternatively, the median tended to overestimate the population median by as much as 10%. It is recommended that usability measurements take a back seat to usability testing when conducting tests with small sample sizes. (PsycINFO Database Copyright 1994 American Psychological Assn, all rights reserved). ************************************************************************

13 OF 118

AU Yanico-Barbara-J. Swanson-Jane-L. Tokar-David-M.

TI A psychometric investigation of the Black Racial Identity Attitude Scale--Form B. Special Issue: Racial identity and vocational behavior.

SO Journal of Vocational Behavior. 1994 Apr Vol 44(2) 218-234.



ID psychometrics of Black Racial Identity Attitudes Scale-Form B & correlation with Version A scoring, African Americans.

AB Examined the psychometric properties of Black Racial Identity Attitudes Scale--Form B (BRIAS-B) through item, subscale, and factor analyses, based on administrations of the scale to 540 African-American Ss. Factor analytic support was found for the Preencounter, Immersion-Emersion, and Internalization subscales. No support was found for an Encounter subscale. However, the factor analysis accounted for a small amount of variance, which, coupled with the finding of marginal internal consistencies, suggested little common variance among the items for the BRIAS-B. Version A and Version B scoring showed little correspondence for some subscales, suggesting the earlier and later scoring versions cannot be considered interchangeable. The Preencounter and Internalization subscales showed highly skewed distributions for a number of items and for subscale scores. (PsycINFO Database Copyright 1994 American Psychological Assn, all rights reserved). ************************************************************************

14 OF 118

AU Dunlap-William-P. Chen-RuSan. Greer-Tammy.

TI Skew reduces test-retest reliability.

SO Journal of Applied Psychology. 1994 Apr Vol 79(2) 310-313.


ID skew & transformation of data, test retest reliability.

AB Demonstrates that skew will reduce the test-retest reliability of otherwise normal data and that the greater the skew is, the greater will be the decrease in reliability. In the first section, equations are derived for the test-retest reliability of lognormal data, a skewed distribution that can be returned to normality by the log transformation. The second analytic treatment derives equations for reliability with power transformations ( x -super(a)) that will produce skew in originally normal data. Finally, analysis of test-retest correlations of actual skewed measures confirmed that if skew was minimized by transformation toward symmetry, then the reliabilities increased; the increase in reliability was related to the extent of skew. Researchers are cautioned to consider transforming data that are skewed to obtain greater symmetry and, thus, improve the test-retest reliability of the resulting measure. (PsycINFO Database Copyright 1994 American Psychological Assn, all rights reserved). ************************************************************************

15 OF 118

AU Dauncey-Karen. Giggs-John. Baker-Karen. Harrison-Glynn.

TI Schizophrenia in Nottingham: Lifelong residential mobility of a cohort.

SO British Journal of Psychiatry. 1993 Nov Vol 163 613-619.



ID lifetime geographical mobility, schizophrenic patients in Nottingham, England.

AB The lifetime geographical mobility of 67 patients suffering from schizophrenia was investigated and related to the ecological structure of the city of Nottingham, UK, where they first presented between 1978 and 1980. Focus was on the processes leading to a threefold rate of schizophrenia in the most socioeconomically deprived area of the city compared with the area of least deprivation. The incidence of schizophrenia correlated closely with levels of social and urban deprivation. Results (1) support a significant relationship between schizophrenia and a tendency to originate in areas of urban deprivation and (2) suggest that the skewed distribution begins early in the lives of the patients and is well established at least 5 yrs before initial contact with the psychiatric services. (PsycINFO Database Copyright 1994 American Psychological Assn, all rights reserved). ************************************************************************

16 OF 118

AU Keselman-H-J. Carriere-Keumhee-C. Lix-Lisa-M.

TI Testing repeated measures hypotheses when covariance matrices are heterogeneous.

SO Journal of Educational Statistics. 1993 Win Vol 18(4) 305-319.


ID Welch-James multivariate approximate degrees of freedom procedure, analysis of unbalanced repeated measures designs with heterogeneous covariance matrices.

AB Shows how a multivariate approximate degrees of freedom procedure based on B. L. Welch (1947, 1951) and G. S. James (1951, 1954), as simplified by S. Johansen (1980), can be applied to the analysis of unbalanced repeated measures designs without assuming covariance homogeneity. Through Monte Carlo methods, it is demonstrated that this approach provides a robust test of the repeated measures main effect hypothesis even when the data are obtained from a skewed distribution. The Welch-James approach also provides a robust test of the interaction effect, provided that the smallest of the unequal group sizes is 5-6 times the number of repeated measurements minus one or provided that a reduced level of significance is employed. (PsycINFO Database Copyright 1994 American Psychological Assn, all rights reserved). ************************************************************************

17 OF 118

AU Adler-Nancy-J.

TI An international perspective on the barriers to the advancement of women managers. Special Issue: Women in management.

SO Applied Psychology: An International Review. 1993 Oct Vol 42(4) 289-300.



ID international perspective on barriers to advancement, female managers.

AB Although the percentage of employed women is increasing in most countries, the number in management remains disproportionately low, and the number holding executive positions remains negligible almost everywhere. This pattern holds across oriental and occidental cultures, across communist, socialist, and capitalist systems, and among economically developed and developing countries. Evidence suggests that the underrepresentation, underutilization, and skewed distribution of female managers is neither coincidental nor random but rather a function of systemic cultural sanctions, educational barriers, legal restrictions, and corporate practices. The myth of women's lack of interest in management is discussed. (French abstract) (PsycINFO Database Copyright 1994 American Psychological Assn, all rights reserved). ************************************************************************

18 OF 118

AU Ogasawara-Haruhiko.

TI / Models of the number of errors using structured parameters in a generalized Poisson distribution and the Polya-Eggenberger distribution.

SO Japanese Journal of Behaviormetrics. 1992 Sep Vol 19(2) 1-13.



ID models of parameters of mixtures of vs generalized Poisson distributions as functions of independent variables, application to errors on intelligence tests, adults.

AB Presents models in which the parameters of generalized Poisson distributions or mixtures of Poisson distributions are functions of the independent variables representing the properties of the studied population. These models are applicable to data on incorrect responses on adult intelligence tests. Numerous statistical equations are presented. (English abstract) (PsycINFO Database Copyright 1994 American Psychological Assn, all rights reserved). ************************************************************************

19 OF 118

AU Ottenbacher-Kenneth-J.

TI The interpretation of averages in health professions research: An empirical examination.

SO Evaluation & the Health Professions. 1993 Sep Vol 16(3) 333-341.



ID interpretation of measures of central tendency & other statistical procedures in health sciences textbooks.

AB A potential misinterpretation regarding measures of central tendency was identified in several health sciences textbooks presenting basic statistical procedures. The misinterpretation involves measures of central tendency derived from skewed unimodal sample distributions. The reviewed textbooks state or imply that in asymmetrical distributions the median is always located between the mode and mean. An example is presented illustrating the fallacy of this assumption. The mean and median will always be to the right of the mode in a positively skewed unimodal distribution and to the left of the mode in a negatively skewed distribution, but the order of the mean and median is impossible to predict or generalize. (PsycINFO Database Copyright 1994 American Psychological Assn, all rights reserved). ************************************************************************

20 OF 118

AU Boiney-Lindsley-G.

TI The effects of skewed probability on decision making under ambiguity.

SO Organizational Behavior & Human Decision Processes. 1993 Oct Vol 56(1) 134-148.



ID skewed probability, ambiguity preference patterns in decision making, MBA students, development of choice model.

AB Argued that the current emphasis on the 1st 2 moments of the 2nd-order distribution overlooks a crucial component of the decision maker's response to ambiguity. The response to nonsymmetric distributions of probability by incorporating the effects of skewness into a theory of preference under ambiguity was examined. Certain effects of skewness on preference were hypothesized. Results of a study exploring the preference patterns of 130 MBA students are presented. A model of preference under ambiguity that could accommodate and explain the observed preference patterns was developed. (PsycINFO Database Copyright 1994 American Psychological Assn, all rights reserved). ************************************************************************

21 OF 118

AU Wilcox-Rand-R.

TI Comparing one-step M-estimators of location when there are more than two groups.

SO Psychometrika. 1993 Mar Vol 58(1) 71-78.


ID modified comparison of 1 step M-estimators of location in psychometric measures with heavy tailed distributions & more than 2 groups.

AB Methods for comparing means are known to be highly nonrobust in terms of Type II errors. The problem is that slight shifts from normal distributions toward heavy-tailed distributions inflate the standard error of the sample mean. In contrast, the standard error of various robust measures of location, such as the 1-step M-estimator, are relatively unaffected by heavy tails. R. R. Wilcox (see PA, Vol 79:29852) examined a method of comparing the 1-step M-estimators of location corresponding to 2 independent groups which provided good control over the probability of a Type I error even for unequal sample sizes, unequal variances, and different shaped distributions. There is an extension of this procedure to pairwise comparisons of more than 2 independent groups, but simulations reported here indicate that it is unsatisfactory. A slight modification of the procedure is found to give much better results. (PsycINFO Database Copyright 1993 American Psychological Assn, all rights reserved). ************************************************************************

22 OF 118

AU Rhee-Kijong.

TI A FORTRAN solution for evaluating the coefficients of the power method for nonnormal transformation.

SO Educational & Psychological Measurement. 1993 Spr Vol 53(1) 107-109.


ID FORTRAN program for identification of coefficients from power method for transformations of standard normal variable to nonnormal variables with prespecified skewness & kurtosis.

AB Presents a FORTRAN program for identifying coefficients derived from the power method developed by A. I. Fleishman (see PA, Vol 63:11110) for nonnormal transformations of a standard normal variable. With this program, researchers can discern whether or not specific nonnormal conditions are attainable. After identifying the appropriate coefficients, researchers can now easily obtain nonnormal variables that have the prespecified skewness and kurtosis from a standard normal variable. (PsycINFO Database Copyright 1993 American Psychological Assn, all rights reserved). ************************************************************************

23 OF 118

AU Bush-Lauren-K. Hess-Ursula. Wolford-George.

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

SO Psychological Bulletin. 1993 May Vol 113(3) 566-579.


ID data transformation, improvement of power & probability of Type I & Type II errors in Monte Carlo simulated within-S design.

AB 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 Copyright 1993 American Psychological Assn, all rights reserved). ************************************************************************

24 OF 118

AU Mortensen-Erik-L. Gade-Anders.

TI Linear versus normalized T scores as standardized neuropsychological test scores.

SO Scandinavian Journal of Psychology. 1992 Sep Vol 33(3) 230-237.



ID use of normalized T scores as standardized neuropsychological test scores, normal vs brain damaged 20-83 yr olds.

AB Both linear and normalized T scores were calculated for 141 normal Ss and a group of 141 patients with diffuse or focal brain damage. All Ss were 20-83 yrs old. Many standard neuropsychological tests have skewed raw score and linear T score distributions, and it is argued that normalized T scores have practical advantages because they permit simple descriptions of both patient groups and individual test score distributions. The authors also argue that skewness can be partially explained by ceiling effects and other test construction artifacts and that skewed raw score distributions do not necessarily reflect skewed distributions of the underlying mental abilities. Consequently, use of normalized T scores seems appropriate in many research and clinical contexts. (PsycINFO Database Copyright 1993 American Psychological Assn, all rights reserved). ************************************************************************

25 OF 118

AU Schifferstein-Hendrik-N. Frijters-Jan-E.

TI Contextual and sequential effects on judgments of sweetness intensity.

SO Perception & Psychophysics. 1992 Sep Vol 52(3) 243-255.



ID skewed sucrose concentration distributions & context & sequential effects, rating of sweetness intensity of sucrose solutions on linear vs 7 point category scale, 20-27 yr olds, Netherlands.

AB Investigated the effect of skewed concentration distributions on ratings on a line scale and compared it to the context effect found for a 7-point category scale. In addition, sequential dependencies between consecutive stimuli and responses were investigated to assess their relevance in taste-intensity scaling studies. The context effects were similar for both scales. Ss were 54 adults. The analyses of sequential effects (SEs) show that both preceding responses and preceding stimuli affect current responses. However, since these 2 factors work in opposite directions, only a small contrast effect was significant. Even though the overall SEs between consecutive stimuli and responses are small, the effect of experimental context may be considerable. It is argued that contextual and SEs are only indirectly related. (PsycINFO Database Copyright 1993 American Psychological Assn, all rights reserved). ************************************************************************

26 OF 118

AU Jansen-Margo-G. Van-Duijn-Marijtje-A.

TI Extensions of Rasch's multiplicative Poisson model.

SO Psychometrika. 1992 Sep Vol 57(3) 405-414.



ID Bayesian approach to extensions of Rasch's multiplicative Poisson model.

AB Considers a model developed by G. Rasch (1960) which assumes that scores observed on some types of attainment tests can be regarded as realizations of a Poisson process. The parameter of the Poisson distribution is assumed to be a product of 2 other parameters pertaining to the ability of the S and to the difficulty of the test. Rasch's model is expanded by assuming a prior distribution with fixed, but unknown, parameters, for the S parameters. The test parameters are considered fixed. How additional between- and within-Ss factors can be incorporated is discussed. Methods for testing the fit and estimating the parameters of the model are discussed and illustrated by empirical examples. (PsycINFO Database Copyright 1993 American Psychological Assn, all rights reserved). ************************************************************************

27 OF 118

AU Kranzler-John-H.

TI The skewness of the distribution of RT trials does not correlate with psychometric g.

SO Personality & Individual Differences. 1992 Aug Vol 13(8) 945-946.



ID correlation of skewness of RT distribution with psychometric g on elementary cognitive tasks, college students.

AB Studied the relationship between the skewness of the distribution of reaction time (RT) trials and psychometric g. Results from 3 different elementary cognitive tasks obtained from 101 college students indicate that RT skewness is not importantly related to g. (PsycINFO Database Copyright 1993 American Psychological Assn, all rights reserved). ************************************************************************

28 OF 118

AU Mellers-Barbara-A. Ordonez-Lisa-D. Birnbaum-Michael-H.

TI A change-of-process theory for contextual effects and preference reversals in risky decision making. Special Issue: Utility measurement.

SO Organizational Behavior & Human Decision Processes. 1992 Aug Vol 52(3) 331-369.



ID context & preference reversal response mode effects, risky decision making in gambling judgment, college students, application of change of process theory.

AB Three experiments were conducted to investigate contextual effects and response mode effects (preference reversals) in risky decision making. Judgments of the worth of binary gambles were examined using 2 different contexts (positively and negatively skewed distributions of expected values) and 2 different response modes (attractiveness ratings and buying prices). Approximately 40 undergraduate Ss participated in each of the 4 conditions. Changes in the response mode affected the preference order of gambles, and changes in the context due to variations in skewing influenced the metric properties of the judgments, but had a minimal effect on preference orders. Results could be described by a change-of-process theory which assumes that the method of elicitation influences the manner in which people combine information and arrive at judgments. (PsycINFO Database Copyright 1993 American Psychological Assn, all rights reserved). ************************************************************************

29 OF 118

AU Wilcox-Rand-R.

TI Comparing the medians of dependent groups.

SO British Journal of Mathematical & Statistical Psychology. 1992 May Vol 45(1) 151-162.



ID comparison method for median of 2 dependent groups.

AB Investigations of the characteristics of real data indicate that psychometric measures can have highly skewed distributions with heavy tails and outliers. In terms of power, the paired t test is known to be unsatisfactory when distributions have heavy tails, and there is the issue of whether some measure of location other than the mean might be more appropriate when distributions are skewed. A method is proposed for comparing the medians of 2 dependent groups. Simulations indicate that the proposed procedure controls the probability of a Type I error to a reasonable degree. In terms of power the new procedure is much less affected by contamination than the paired t test or the Wilcoxon signed rank test. (PsycINFO Database Copyright 1992 American Psychological Assn, all rights reserved). ************************************************************************

30 OF 118

AU Wilcox-Rand-R.

TI Comparing one-step M-estimators of location corresponding to two independent groups.

SO Psychometrika. 1992 Mar Vol 57(1) 141-154.


ID comparison of 1 step M-estimates of location in psychometric measures with heavy tailed distributions.

AB Suggests an approach, based on the 1-step M-estimator of location, to be used when psychometric measures have heavy-tailed distributions. Simulations indicate that the procedure provides good control over the probability of a Type I error even when the distributions are skewed, have different shapes, and the variances are unequal. The procedure has more power than B. Welch's (1937) method when distributions have heavy tails, and it compares well to K. K. Yuen's (1974) method for comparing trimmed means. R. R. Wilcox's (1990) median procedure has about the same power as the proposed procedure, but Wilcox's method is based on a statistic that has a finite sample breakdown point of only 1/ n, where n is the sample size. (PsycINFO Database Copyright 1992 American Psychological Assn, all rights reserved). ************************************************************************