Explore chapters and articles related to this topic
Nonparametric Methods
Published in Marcello Pagano, Kimberlee Gauvreau, Heather Mattie, Principles of Biostatistics, 2022
Marcello Pagano, Kimberlee Gauvreau, Heather Mattie
The sign test may be used to compare two populations which are not independent. In this respect, it is similar to the paired t test. A random sample of paired observations is selected from the two populations of interest. The test then focuses on the difference in values within each pair. However, it does not require that the population of differences be normally distributed. Furthermore, whereas the null hypothesis of the paired t test is that the mean of the underlying population of differences is equal to 0, the null hypothesis of the sign test is that the median difference is equal to 0.
Non-Parametric Tests of Significance
Published in M. Venkataswamy Reddy, Statistical Methods in Psychiatry Research and SPSS, 2019
The sign test is used when the researcher involves in hypothesis that there is change after the treatment. This test is applied when the data on a variable have underlying continuity but which can be measured in only a very gross way.
Non Parametric Methods
Published in Marcello Pagano, Kimberlee Gauvreau, Principles of Biostatistics, 2018
Marcello Pagano, Kimberlee Gauvreau
The sign test may be used to compare two samples of observations when the popula‐ tions from which the values are drawn are not independent. In this respect, it is similar to the paired t‐test. Like the t‐test, it does not examine the two groups individually; in‐ stead, it focuses on the difference in values for each pair. However, it does not require that the population of differences be normally distributed. The sign test is used to eval‐ uate the null hypothesis that in the underlying population of differences among pairs, the median difference is equal to 0.
Pre-employment hearing threshold levels of 59,601 Australian male coal miners compared to an otologically normal international male population (ISO7029:2019)
Published in International Journal of Audiology, 2023
Adelle Liebenberg, Valerie M. Nie, Alan M. Brichta, Sima Ahmadi, Carole L. James
The differences in median HTL between the mining population and the ISO population HTL were analysed using the sign test. Sign tests were conducted by cohort, age, and frequency, testing the null hypothesis that the distribution of the observed differences between the median HTLs of the mining population and the ISO population, for each cohort, age, and frequency category, has a median equal to zero. The sign-test calculates the difference between the populations, without assuming that the mining population HTL deviation is higher or lower than the ISO population. The unadjusted p-values for the sign test results, using the Bonferroni threshold of p < 0.0005, are reported in Table 2. The results are presented for each age category and cohort separately, in order to describe the HTL distributions across the cohorts and allow comparison to the ISO population.
Intervention in an opioid overdose event increases interest in treatment among individuals with opioid use disorder
Published in Substance Abuse, 2021
Jermaine D. Jones, Aimee N. Campbell, Laura Brandt, Felipe Castillo, Rebecca Abbott, Sandra D. Comer
The current analysis sought to assess whether intervention in an overdose event led to changes in the “importance” of substance use treatment. For the current analysis, overdose “intervention” was defined as the administration of naloxone. Data were analyzed for all participants who reported intervention in an overdose event, regardless of their duration of participation in the trial. For participants who intervened in multiple overdoses, investigators only assessed change in treatment importance following their first overdose intervention, while enrolled in the trial. The Shapiro–Wilk test was used to determine if the data were normally distributed.21 In cases where parametric tests were appropriate, t-tests were used to compare mean ratings at the time point most immediately before and after participants intervened in an overdose event. In cases of a significant Shapiro-Wilkes test, the non-parametric Sign Test was used.22 Repeated-measures analysis of variance (ANOVA) or the non-parametric Friedman test was used to compare treatment importance at each of the follow-up visits. All hypothesis tests were two‐sided and the significance level was set at p < 0.05. Analyses were performed in SAS® 9.4 (SAS Institute Inc.) and SPSS Version 25 (IBM).23,24
Health-related quality of life of long-term advanced melanoma survivors treated with anti-CTLA-4 immune checkpoint inhibition compared to matched controls
Published in Acta Oncologica, 2021
A. H. Boekhout, A. Rogiers, K. Jozwiak, M. J. Boers-Sonderen, A. J. van den Eertwegh, G. A. Hospers, J. W. B. de Groot, M. J. B. Aarts, E. Kapiteijn, A. J. ten Tije, D. Piersma, G. Vreugdenhil, A. A. van der Veldt, K. P. M. Suijkerbuijk, E. A. Rozeman, B. Neyns, K. J. Janssen, L. V. van de Poll-Franse, C. U. Blank
Differences in EORTC QLQ-C30 and MFI scores between survivors and controls were estimated and adjusted for age, gender, education (primary/high school/vocational vs. college/university), and marital status (not partnered vs. partnered) using generalized estimating equations with clustering together matched individuals. Spearman’s correlation coefficients were used to assess associations between functioning and symptom burden, EORTC QLQ-C30 scores and cumulative ipilimumab dose, and MFI scores and demographic characteristics. Mann–Whitney or Kruskal–Wallis tests were used to compare EORTC QLQ-C30 scores between clinical, treatment characteristics and HADS scores. The Mann–Whitney test was also used for comparing HRQoL outcomes between survivors with vs. without brain metastases, survivors with a follow-up <36 vs. ≥36 months post ipilimumab treatment and survivors with ≥2 vs. without comorbidities. Sign test was used for comparison of questionnaire evaluated at two-time points. Missing items from the EORTC QLQ-C30 were imputed according to EORTC guidelines. Of all questionnaires, the scale score was set to missing, if fewer than half of the items on a given scale were answered. Since our research was a hypothesis-generating research, none of the p values were corrected for multiple testing.