The power of statistical tests in meta-analysis.

Calculations of the power of statistical tests are important in planning research studies (including meta-analyses) and in interpreting situations in which a result has not proven to be statistically significant. The authors describe procedures to compute statistical power of fixed- and random-effects tests of the mean effect size, tests for heterogeneity (or variation) of effect size parameters across studies, and tests for contrasts among effect sizes of different studies. Examples are given using 2 published meta-analyses. The examples illustrate that statistical power is not always high in meta-analysis. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Color vision tests for early detection of antiepileptic drug toxicity

A previous suggestion that antiepileptic drugs may induce color vision deficiencies prompted us to examine whether color vision deficiencies may occur at lower drug serum concentrations than those associated with symptoms of neurotoxicity. Eighty patients presenting with epilepsy received monotherapies of valproic acid, phenytoin, or carbamazepine; 18 patients did not receive antiepileptic drug therapy. Color vision was tested by the Farnsworth-Munsell 100-hue test, spectral sensitivity, and the newly developed tritan screening plates. Patients treated with phenytoin or carbamazepine developed blue-yellow color vision deficiencies. In contrast, patients exposed to valproic acid or receiving no drug treatment showed normal color vision. There was a significant correlation (p < 0.0001) between signs of neurotoxicity induced by phenytoin or carbamazepine and blue-yellow color vision deficiencies. In contrast, we found no correlation between these signs of neurotoxicity and the drug serum concentrations (p = 0.0637). Color vision testing in epileptic patients treated with phenytoin or carbamazepine appears to be a sensitive method for early detection and monitoring of clinical neurotoxicity.     

Approximating power for significance tests with one degree of freedom.

Shortcut approximate equations are described that provide estimates of the sample size required for 50% power (α?=?0.05, two-tailed) for 1 degree of freedom tests of significance for simple correlations, differences between 2 independent group means, and Pearson's chi-square test for 2?×?2 contingency tables. These sample sizes should be thought of as minima, because power equal to 50% means that the chance of a significant finding is that of flipping a fair coin. A more desirable sample size can be computed by simply doubling the 50% sample sizes, which is shown to result in power between 80% and 90%. With these simple tools, power can be estimated rapidly, which, it is hoped, will lead to greater use and understanding of power in the teaching of statistics and in research. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Strategies for making friends: Relations to social behavior and peer acceptance in early adolescence.

Examined relations among strategy knowledge about making friends, prosocial and antisocial behavior, and peer acceptance at school during early adolescence. Based on a sample of 423 6th- and 7th-grade students, findings indicated that knowledge of both appropriate and inappropriate strategies for making friends was related significantly to both types of social behavior and to peer acceptance. Results also suggested that displays of prosocial (but not antisocial) behavior represent an intermediate process that links knowledge about making friends to peer acceptance. These results were obtained even when taking into account significant relations of IQ to strategy knowledge, social behavior, and peer acceptance. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A power study of a sequential method of p-value adjustment for correlated continuous endpoints

In the course of clinical (or preclinical) trial studies, it is a common practice to conduct a relatively large number of tests to extract the maximum level of information from the study. It has been known that as the number of tests (or endpoints) increases, the probability of falsely rejecting at least one hypothesis also increases. Single-step methods such as the Bonferroni, Sidák, or James approximation procedure have been used to adjust the p-values for each hypothesis. To reduce the conservatism (i.e., underestimating type I error) possessed by the aforementioned methods, Holm proposed a so-called "free-step-down" procedure. This adjustment can be made even less conservative by incorporating the dependence structure of endpoints at each adjustment step of the procedure. That is done by sequentially applying James's approximation procedure for correlated endpoints at each step, referred to as the Free-James method. This article primarily compares the power of the Free-James method to the power of the Bonferroni and James single-step-down and the Holm free-step-down methods. Two definitions of power are considered: (a) the probability of correctly rejecting at least one hypothesis when it is true, and (b) the probability of correctly rejecting all hypotheses that are true. Monte Carlo simulations show that the Free-James method is as good as other methods under definition (a) and the most powerful under definition (b) for various sample sizes, numbers of endpoints, and correlations.     

A longitudinal assessment of the academic correlates of early peer acceptance and rejection

Examines the extent to which academic achievement and work habits of first and second graders are predicted by classroom social status over the kindergarten, first-, and second-grade period. Three hundred and forty five children (163 boys and 182 girls) from a southern California community comprised the sample. The ethnic distribution of the sample was approximately 45% Caucasian, 42% Latino, 9% African American, and 5% Asian or other ethnicity. Findings suggest that peer rejection assessed as early as kindergarten and social rejection that is stable across 2 years (kindergarten-first grade or first-second grade) are associated with deficits in first-grade work habits and second-grade academic achievement and work habits. In contrast, stable social acceptance appears to buffer children from early academic difficulty. The pattern of findings remain significant after controlling for initial kindergarten academic competence. The implications for clinical and educational intervention programs are discussed.     

Size and power of two-sample tests of repeated measures data

One method of using repeated measures data to compare treatment groups in a clinical trial is to summarize each subject's outcomes with a single summary statistic, and then perform a distribution-free comparison based on the resulting statistics. We examine extensions of this approach and conditions under which they retain proper size in the presence of missing data. The asymptotic relative efficiencies of several summary statistic tests are calculated to show which perform best in a variety of situations. The techniques are illustrated using data from an AIDS clinical trial.     

Evaluation of serological tests for the detection of pseudorabies gE antibodies during early infection

Two enzyme-linked immunosorbent assays (ELISAs) and a particle concentration fluorescence immunoassay (PCFIA) were compared for their ability to detect antibodies against pseudorabies virus (Aujeszky's disease virus) glycoprotein E (gE) in the early stages of infection in pigs previously vaccinated with gE-deleted pseudorabies vaccines. Seventy pigs were included in the study. Five groups of 6 pigs each were vaccinated with one of 5 different pseudorabies virus (PRV) gE-deleted vaccines, and subsequently infected intranasally with 10(5.6) TCID50 of the Iowa 4892 pneumotropic strain of PRV. This entire procedure was repeated using 10(4.6) TCID50 of the Rice strain of PRV. Five unvaccinated control pigs were also challenged with each virus strain. Three control pigs died before seroconverting, leaving 67 pigs for comparison. Blood samples were drawn from experimentally inoculated pigs on the day of vaccination, the day of challenge, and on 4-10, 14, and 21 days postchallenge (DPC). Serology test sensitivity estimates and comparisons among tests were made for each sampling day. Results of this study demonstrated differences among the tests in the time from inoculation to initial antibody detection, and the time to detect 50% and 75% of the infected pigs. The average time until first detection of pigs as seropositive for gE antibodies by PCFIA was 7.5 DPC. The blocking ELISA detected pigs as seropositive an average of 8.8 DPC, and the indirect ELISA first detected gE antibodies by 9.3 DPC. Fifty percent of the pigs were detected as seropositive by days 7, 8, and 9 for the PCFIA, blocking ELISA, and indirect ELISA, respectively. Similarly, 75% of the pigs were detected as seropositive by days 8, 9, and 10 for the PCFIA, blocking ELISA, and indirect ELISA, respectively. All pigs were detected as seropositive by 14 DPC for all 3 tests.     

Dependent variable reliability and the power of significance tests.

Assuming that the linear models for classical test theory and ANOVA hold simultaneously for some dependent variable, it is shown that 2 contradictory statements concerning the relationship between reliability and statistical power are both correct. J. E. Overall and J. A. Woodward (see PA, Vol 53:8623, 57:7284) showed that when the reliability of a difference or change score is zero, the power of a statistical test of a hypothesis of no change can be at a maximum. J. L. Fleiss (see record 1977-07259-001) found the opposite result (i.e., that the power of a statistical test of no pre–post change is at a maximum when the reliability of the difference or gain scores is equal to one). The role of the reliability of the dependent variable in statistical evaluations of controlled experiments is examined. It is argued that the conditions that yield high reliability coefficients are not necessarily optimal for significance testing. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Effect of error variance heterogeneity on the power of tests for regression slope differences.

In moderated regression analysis with both a continuous predictor and nominal-level (group membership) variables, there are conditions in which the hypothesis of equal slopes of the regression of Y onto X across groups is equivalent to the hypothesis of equality of X–Y correlations across groups. This research uses those conditions to investgate the impact of heterogeneity of error variance on the power accuracy of the F test for equality of regression slopes. The results show that even when sample sizes are equal, the test is not robust and, under unequal sample sizes, the pattern of excessively high or excessively low rejection rates can be severe. (PsycINFO Database Record (c) 2011 APA, all rights reserved)     

电站锅炉用TP347H钢管爆裂分析

利用金相显微镜和扫描电镜等分析技术,分析电站锅炉用TP347H钢管的爆裂原因.结果表明,钢管中碳化物在高温下沿晶界聚集分布,会降低抗晶间腐蚀能力,材料的蠕变抗力下降,并在钢管表面产生沿晶裂纹,在高温高压作用下裂纹不断扩展并最终产生爆管.     

Seventy-two tests of the sequential lineup superiority effect: A meta-analysis and policy discussion.

A decade ago, a meta-analysis showed that identification of a suspect from a sequential lineup versus a simultaneous lineup was more diagnostic of guilt (Steblay, Dysart, Fulero, & Lindsay, 2001). Since then, controversy and debate regarding sequential superiority has emerged. We report the results of a new meta-analysis involving 72 tests of simultaneous and sequential lineups from 23 different labs involving 13,143 participant-witnesses. The results are very similar to the 2001 results in showing that the sequential lineup is less likely to result in an identification of the suspect, but also more diagnostic of guilt than is the simultaneous lineup. An examination of the full diagnostic design dataset (27 tests that used the full simultaneous/sequential × culprit-present/culprit-absent design) showed that the average gap in correct identifications favoring the simultaneous lineup over the sequential lineup—8%—is smaller than the 15% figure obtained from the 2001 meta-analysis (and from the current full 72-test dataset). The lower error rate incurred for culprit-absent lineups with use of a sequential format remains consistent across the years, with 22% fewer errors than simultaneous lineups. A Bayesian analysis shows that the posterior probability of guilt following an identification of the suspect is higher for the sequential lineup across the entire base rate for culprit presence/absence. New ways to think about policy issues are discussed. (PsycINFO Database Record (c) 2011 APA, all rights reserved)