Two-group Test Research Articles

Smoking reduction among high school seniors: a test of selected indicators.

To examine differences in the explanatory power of variables thought to discriminate between students who smoke and students who have reduced (or quit). The design is a two-group test of differences on 33 variables identified from previous studies of adolescent smoking and cessation as potential discriminators, which represent a variety of perceptual, behavioral, social, and psychological factors. Data are from a national survey of U.S. high school students administered in 1997 by the Monitoring the Future project. The most effective variables involve smoking-related beliefs and perceptions, particularly in anticipated future smoking status, attempted quitting, number of smoking friends, and perceptions of other smokers. Personality and psychological variables were not effective at discriminating between the two groups.

TYPE I ERROR CONTROL OF TWO-GROUP MULTIVARIATE TESTS ON MEANS UNDER CONDITIONS OF HETEROGENEOUS CORRELATION STRUCTURE AND VARIED MULTIVARIATE DISTRIBUTIONS

ABSTRACT A Monte Carlo study was conducted to evaluate the performance of two-group multivariate tests on means under conditions of homogeneous and heterogeneous correlation structure and multivariate normality and nonnormality. The test statistics under study included those based on the work of Hotelling and alternative procedures proposed by James, Yao, Johansen, Nel and van der Merwe, and Kim. Results indicated that the F statistic based on Hotelling's T 2 was outperformed by the alternative procedures which showed good Type I error control under moderate to large sample sizes. Under small sample sizes, the alternative procedures tended to have unacceptable Type I error control. The F statistic based on Hotelling's T 2 is recommended for use when the sample size in each group is small (e.g., n = p + 1) and correlation heterogeneity is mild (<.3), or when groups are approximately equal in size. Though simulated conditions of multivariate nonnormality had some impact on the Type I error rates on some of the procedures, the impact was in general small and apparent only under the smaller sample size conditions.

An Age-Adjusted Trend Test for the Tumor Incidence Rate for Multiple- Sacrifice Experiments

The nonparametric age-adjusted trend test for the tumor incidence rate proposed by Malani and Lu (1993, Communications in Statistics--Theory and Methods 22, 1557-1584) is modified in two ways to improve the size of the test. First, the tumor incidence rate is constrained to be nonnegative, and second, a conditional variance is used in place of the asymptotic variance based on the delta method. A Monte Carlo simulation study demonstrates the improvement in size of test. The proposed test can be viewed also as a modification of the two-group test of Dewanji and Kalbfleisch (1986, Biometrics 42, 325-341), although their maximum likelihood routine is based on a different formulation of the problem.

An Age-Adjusted Trend Test for the Tumor Incidence Rate for Single-Sacrifice Experiments

The nonparametric age-adjusted trend test for the tumor incidence rate proposed by Malani and Lu (1993, Communications in Statistics--Theory and Methods 22, 1557-1584) is modified in two ways to improve the size of the test. First, the tumor incidence rate is constrained to be nonnegative, and second, a conditional variance is used in place of the asymptotic variance based on the delta method. A Monte Carlo simulation study demonstrates the improvement in size of test. The proposed test can be viewed also as a modification of the two-group test of Dewanji and Kalbfleisch (1986, Biometrics 42, 325-341), although their maximum likelihood routine is based on a different formulation of the problem.

A Monte-Carlo Study of Type I Error Rates and Power for Tukey's Pocket Test

Abstract Tukey's pocket test (1959) is a two-group test of mean differences for which the calculations do not require a calculator or a book of tables. The critical values for the test were derived mathematically by Tukey, and the Type I errors and power of the test when n = 20 and n = 40 for normal and bimodal distributions were studied by Neave and Granger (1968). As an extension of the findings of Neave and Granger, a Monte-Carlo simulation to study Type I errors and power for Tukey's pocket test, with sample sizes ranging from 5 to 100 for both normal and skewed distribution shapes, was conducted. The critical values derived by Tukey worked well, even if the data were heavily skewed. With normal data, the test was reasonably powerful for smaller sample sizes but lost relative power for larger sample sizes, compared with the t test. With skewed data, Tukey's pocket test was often more powerful than the t test.

A single-blind comparison of topical erythromycin/zinc lotion and oral minocycline in the treatment of acne vulgaris

This single-blind multicentre study compared the efficacy of topical 4% erythromycin and 1.2% zinc complex (Zineryt) with 50 mg twice daily oral minocycline in patients with acne vulgaris over a 12-week period. Patients' acne improved on both treatments. The erythromycin/zinc complex was more effective in reducing the numbers of comedones by 4 weeks (P = 0.01, Wilcoxon's two-group test for unpaired data) and at 12 weeks (P = 0.004) and superficial inflamed lesions by 12 weeks (P = 0.02). There was no difference in the speed of onset of response to treatment and both treatments were well tolerated.