China
Africa
Explore chapters and articles related to this topic
Non-Parametric Tests of Significance
Published in M. Venkataswamy Reddy, Statistical Methods in Psychiatry Research and SPSS, 2019
The median test may be used when the null hypothesis states that the two groups are from population with the same median. This method is applicable whenever the scores of the two groups are in at least an ordinal scale. To perform the median test, we first determine the median score of the combined group. Then, we dichotomize both sets of scores at that combined median. Then, the procedure is to prepare a 2×2 contingency table and apply the chi-square test of significance.
Empirical Likelihood
Published in Albert Vexler, Alan D. Hutson, Xiwei Chen, Statistical Testing Strategies in the Health Sciences, 2017
Albert Vexler, Alan D. Hutson, Xiwei Chen
When developing or extending a median test for censored cases, one may face two questions. One is the way to handle the ambiguous observations, that is, the censored observations that are less than the pooled-sample median. The other is the definition of the pooled median. The empirical likelihood approach may answer these questions naturally. Naik-Nimbalkar and Rajarshi (1997) adopted the empirical likelihood approach for testing the equality of k medians based on censored data. This test does not require the assumption of the underlying k survival functions being identical under the null hypothesis, which is common in most of the existing tests. The author supplemented some details to the proof of Thomas and Grunkemeier (1975) and proved that the asymptotic distribution of the test statistic is chi-square with k – 1 degrees of freedom, when the null hypothesis of equality of k median holds.
A comparison of sampling approaches for monitoring schoolwide inclusion program fidelity
Published in International Journal of Developmental Disabilities, 2020
H1 asked if there were significant differences in agreement between expert raters and staff when using the IPRT for each of three types of data collection approaches. The average scores on the IPRT were: proportional sample, n = 69, x̅ = 2.85, SD = .56; proportional sample expert n = 1, x̅ = 3.09; random sample n = 7, x̅ = 3.34, SD = .36; random sample expert rater n = 1, x̅ = 3.00; consensus sample (consensus score of six people) n = 1, x̅ = 2.97; and consensus expert n = 1, x̅ = 3.00. Overall scores and item-level comparisons were conducted for the expert and the sample at the item level for both the proportional and random sample groups. Due to the low sample size, the Median Test was conducted using only the total rating for the consensus group. Non-parametric statistics are appropriate to analyze data with small sample sizes (Siegel and Castellan 1988). None of the differences between the expert and the samples were significantly different for any of the sampling approaches (proportional sample p < .109, random sample p < .063, consensus sample p < .317). While the Median Test does not provide the ability to measure the power of the significance, it does allow researchers to determine if there are any underlying differences in the distribution of the data (Siegel and Castellan 1988).
Rate of RhD-alloimmunization after the transfusion of RhD-positive red blood cell containing products among injured patients of childbearing age: single center experience and narrative literature review
Published in Hematology, 2021
Mark Yazer, Darrell Triulzi, Jason Sperry, Alain Corcos, Jansen Seheult
A Chi-squared test or Fisher's exact test, where appropriate, was used to compare the differences between categorical/dichotomous variables. A median test was performed to determine the likelihood that two or more independent samples came from populations with the same median. Royston's trend test for proportions was performed to determine if there was a significant trend in the RhD-alloimmunization rate across the strata of the number of RhD-positive units transfused. Data analysis was performed in Stata version 16 (Statacorp, TX). P-values were not adjusted for multiple comparisons. This protocol was approved by the University of Pittsburgh's Institutional Review Board.
Dissemination of Information on Neutropenic Diet by Top US Cancer Centers: In-line with the Evidence?
Published in Nutrition and Cancer, 2019
Timothy J. Brown, Dhruvika Mukhija, Naveen Premnath, Anand Venkatraman, Sajan Jiv Singh Nagpal, Arjun Gupta
The age of the website (as calculated from the difference between 02/12/2018 and the website update date) is presented using descriptive statistics. The US news ranks for the individual websites (1–20) were converted to two categories (top 10 and bottom 10) and the recommendation made by the websites (i.e. for, against, or not applicable) were then compared using Fischer-Exact test (non-parametric) considering the small number of responses in each category. Significance was set at p < 0.05. The Median test was used for comparison of median values where applicable. All analyses were performed using an institutionally licensed copy of JMP (SAS Institute, Cary, NC)