Search

Shortly after my retirement, I became involved in local water issues and learned that I could be of especial use in work on those issues because of my background in statistics. What I am reporting here is a minimally-technical account of the experience I had using statistics as a forensic tool to unearth what appears to have been a misuse of statistics in the modeling of the environmental impact of a proposed desalination project to provide water for the community where I reside. Of potential interest to statisticians is how statistics can be useful in detective work. [ABSTRACT FROM AUTHOR]

In an ideal multiple-choice test, defined as a multiple-choice test containing only items with options that are all equally guessworthy, the probability of guessing the correct answer to an item is equal to the reciprocal of the number of the item's options. This article presents an asymptotically exact estimator of the test-retest reliability of an ideal multiple-choice test. When all test items have the same number of options, computation of the estimator requires, in addition to the number of options per item, the same information as computation of the Kuder-Richardson Formula 21: the total number of items answered correctly on a single testing occasion by each person tested. Both for ideal multiple-choice tests and for nonideal multiple-choice tests for which the averageprobability of guessing the correct answer to an item is equal to the reciprocal of the number of options per item, Monte Carlo data show that the estimator is considerably more accurate than the Kuder-Richardson Formula 21 and, in fact, is very nearly exact in populations of the order of 1000 persons.

Psychologists who continue to test Hodespite repeated rejections must be getting what they want. What psychologists get from the test may not be what psychology needs, however, as a number of critics of the test have pointed out. Using as an illustration an experiment singled out in the literature for its methodological excellence, this article not only adds to the criticism by citing seven specific pitfalls likely to trap users of the test but also investigates the motivational forces sustaining the continued use of the test in an effort to overcome the resistance to change.

The 1978 Uniform Guidelines on Employee Selection Procedurescontains three different definitions of unfairness in selection without acknowledging the differences. The fairness counterparts of these definitions correspond to Definitions 1, 2, and 3 of test fairness described in 1971 by Darlington. Suggestive of quotas, Definition 2 no longer commands attention among test users. A test is fair according to Definition 1 if people having equal test scores tend also to be equal in criterion performance while, according to Definition 3, a test is fair if people equal in criterion performance tend also to have equal test scores, regardless of how these people might differ demographically. Satisfaction of Definition 1 yields higher predictive validity than satisfaction of any other definition of test fairness. The difference in the case of Definition 3 tends in practice to be equal to only .02, however, and Definition 3 appears to be the more logically defensible of the two definitions. Although on tests that satisfy either definition of fairness black examinees generally score lower than white examinees, black and white examinees will tend to score four-fifths of a standard-deviation unit closer on a test that satisfies Definition 3 than on a test that satisfies Definition 1. Users of the 1978 Guidelines should be aware of these differences.

Abstract: In this method of analyzing learning data, entire learning curves are described quantitatively by single numbers which are used in a statistical test to determine whether two or more groups of learning curves are significantly different. The method has some logical advantages over prevailing methods in that it avoids the use of average learning curves and of arbitrary measures of slope and asymptote. Its disadvantage is computational. Since it involves the use of factor analytic procedures, it may be tedious to apply unless computation is carried out on a high-speed computer.

HH is a genetic disorder commonly associated with homozygosity for the C282Y HFEmutation and characterized by progressive iron overload through increased intestinal absorption. Organ failure due to iron toxicity may develop. Iron removal by phlebotomy is the preferred treatment and has been demonstrated to prevent or reverse some of the complications of iron overload. However, compliance with a weekly phlebotomy schedule is variable, and some patients are ineligible for phlebotomy due to underlying medical disorders. Thus, if an oral iron chelator such as deferasirox proves to be safe and effective, HH patients will have an alternative treatment option. This is an inter-patient dose-escalation study of deferasirox (5, 10, 15, 20 mg/kg) given daily for 24 weeks to C282Y homozygous HH patients with a pre-treatment serum ferritin (SF) value ≥300 μg/L and ≤2000 μg/L, and transferrin saturation ≥45%. Major exclusion criteria are men with hemoglobin <13 g/dL, women with hemoglobin <12 g/dL, a history of blood transfusion during 6 months prior to study entry, serum creatinine above the upper limit of normal (ULN), and serum ALT ≥2xULN at screening. The primary endpoint is the incidence and severity of adverse events (AEs). Secondary endpoints include the change in SF from baseline at 24 weeks, the time to normalization of SF (defined as the first occurrence of reduction of SF to <100 μg/L), the longitudinal course of SF, and the pharmacokinetics of deferasirox. It is estimated that at least 40 patients are needed to evaluate safety at all dose levels. Cohorts of 8 patients per dose level will be used in order to detect AEs with a 25% true incidence rate at that dose with 90% power. Safety monitoring will be based on medical review and a 2-parameter Bayesian logistic regression model for dose-dependent probabilities of a severe AE. To assess efficacy, the change from baseline in SF after 24 weeks of treatment will be analyzed by performing an analysis of covariance (ANCOVA). To date, 11 patients (9 men, 2 women; all Caucasian; mean age 56 years) with a mean of 7 years since HH diagnosis have been treated at 5 mg/kg/day for at least 4 weeks. There was a mean of 7 years since HH diagnosis, with 2 patients not having been previously treated. The remaining 9 had been treated with phlebotomy, one of whom had also been treated with deferoxamine. Baseline iron studies and ALT values for the 11 patients treated at 5 mg/kg/day are summarized in the table. The dose of deferasirox has been escalated to 10 mg/kg/day after no patients were seen to experience severe AEs at 5 mg/kg/day. In conclusion, this ongoing study will generate preliminary safety and efficacy data for deferasirox use in iron-overloaded HH patients, indicating whether deferasirox could be an alternative to phlebotomy in selected patients.

The newborn is capable of secreting AVP, but there are no data regarding relative responsiveness to the several known physiological stimuli. We examined AVP, plasma sodium (Ma) and osmolality (Osm) in response to phlebotomy, water loading, hypertonic saline and mild dehydration in 7-11 lambs, 2-49 (mean = 19) days of age. AVP was measured by radioimmunoassay.Repeated phlebotomies (total 20 ml/kg) raised AVP from 4.8 ± 2.1 (mean ± SEM) to 74 ± 19 μU/ml (p < .01), while Na and Osm remained unchanged. When 100 ml/kg 2.5% dextrose/water was infused over 60 minutes, AVP fell from 3.,4 ± 1.2 to 0.7 ± 0.6 μU/ml by 60 minutes (p < .05). Na and Osm fell from 140 to 125 mEq/L and 283 to 262 mOsm/kg respectively (p < .05) Hypertonic (23%) sodium chloride infusion (10 mEq/kg) increased Na from 142 to 159 mEq/L and Osm from 271 to 318 mOsm/kg over a 30 min period (p < .05). In response to this stimulus, AVP increased from 2.9 ± 0.7 to 22.2 ± 9 μU/ml (p < .05). After 18 hours of dehydration, AVP rose from 0 6 ± 0.1 to 34 ± 1 8 μU/ml, Na from 134 to 140 mEq/L (p < 05) and Osm from 293 to 306 mOsm/kg (p < .05).Thus the newborn lamb is capable of responding to both volume and osmolar stimuli. The quantitative stimulus-response ratios (SRR = Mog AVP/ΔOsm) were similar for water loaded and saline stimulated newborns and similar to responses in the adult. The dehydration SRR was not accountable by osmolar change alone but also reflected volume change.

Cord blood levels of arginine vasopressin(AVP)are known to be high after vaginal delivery(VD)in contrast to ceasarian section (C/S)delivery without labor. However, the mechanism, significance, and postnatal duration of the AVP hypersecretion are not known. We studied serial values for AVP concentration by radioimmunoassay in 61 full term and 13 premature (32-36 weeks gestation) infants during the first 24 hours of life. All normal infants had Apgar values >7. Mean AVP values vs time did not differ for premature and full term infants. Paired (mean ± SEM, uU/ml) umbilical artery(UA)and vein(UV)levels following VD and C/S during labor were 90 ± 15 and 15 ± 5.6 respectively(p<.001). Cord UA values for 2 anencephalic infants, 2 infants with breech presentation, and 9 infants delivered by C/S without labor, were significantly lower than values after normal VD and C/S with labor. Plasma AVP after VD or C/S with labor fell rapidly after delivery; peripheral venous levels in VD infants were 22% of UV values by 30 minutes (p<.05). By 2 hours AVP values for all infants approached adult basal levels.Conclusion: a)Cord blood AVP is of fetal origin since UA>UV, and anencephalic AVP levels are low, b)Premature infants have an intact AVP secretory mechanism, c)Anoxia is not the stimulus for the fetal AVP surge in labor, d)Cerebral compression from VD may be the mechanism of AVP hypersecretion.

For 1,610 classes, mutually exclusive contributions of faculty and course to over-all evaluation of instruction were 45% and 15%, respectively, the remainder of the response variation being due to interaction and error.

Simultaneous maternal and fetal plasma oxytocin (OXY) concn. (µU/ml) were measured by radioimmunoassay before and during continuous infusion of synthetic OXY to steady state conditions into ewe or fetus (gestational age 124-140 days; at least 5 days post-surgery; estimated fetal wt. 3 kg).Fetal metabolic clearance rates (MCR in ml/kg/min) were calculated to be 18.1 ± 1.1 and 15.0 ± 1.3 at the two infusion rates; maternal MCR were 12.7 ± 2.8 and 13.4 ± 2.3, respectively. Examination of simultaneous fetal and maternal baseline OXY concentrations revealed that fetal levels were significantly higher than maternal: 1.9 ± 0.2 vs 0.7 ± 0.1 (p<.05). Continuous monitoring of uterine pressure revealed that uterine contractions were induced by maternal infusion of 800 µU/kg/min; no uterine contractions were induced by fetal infusion. Conclusions based on these data are: 1) plasma OXY levels exceed maternal levels in fetuses of 124-140 days gestation; 2) transplacental passage of OXY is minimal in both M-F and F-M directions; 3) maternal and fetal MCR of OXY are similar and unrelated to plasma OXY levels.