China
Africa
Explore chapters and articles related to this topic
The Basics of Statistical Tests
Published in Mitchell G. Maltenfort, Camilo Restrepo, Antonia F. Chen, Statistical Reasoning for Surgeons, 2020
Mitchell G. Maltenfort, Camilo Restrepo, Antonia F. Chen
A statistic does not just have to summarize a single variable. A statistical test summarizes the association between measurements of interest (such as treatment and outcome) in a single number called a test statistic. This test statistic is compared against a known distribution, which might not be Gaussian, and the p-value estimates that probability for which a value of at least that observed test statistic would have happened, if it is true that the measurements of interest are in fact not related. This assumption is called the null hypothesis since it can be generally stated as the assumption that the expected difference between the two groups is zero (that is, null). Expressed another way, if two groups have the same distribution, then the null hypothesis is that the expected measurement from one is the same as the expected measurement from another. (If the two groups do not have the same distribution – perhaps one has a wider variance or is more skewed – then there’s something going on you should investigate.) There is also a variation called a sharp null hypothesis, which assumes for every individual in a group, the expected effect of a treatment is zero; this is different from the expectation that some patients may get better and some worse, so the expected average effect is zero.
THE ANALYSIS OF ANIMAL CARCINOGENICITY EXPERIMENTS
Published in Richard G. Cornell, Statistical Methods for Cancer Studies, 2020
Richard G. Cornell, Robert A. Wolfe, William J. Butler
Here it suffices to emphasize that it is particularly important to achieve high power, that is, high probability of detection of an effect if it exists. This is facilitated by the use of one-sided tests, which is appropriate in any test of toxicity, and by the incorporation of evidence from different strata into a single test statistic or confidence interval, as just suggested.
Meta-Analysis of Dose-Response Relationships
Published in Christopher H. Schmid, Theo Stijnen, Ian R. White, Handbook of Meta-Analysis, 2020
Nicola Orsini, Donna Spiegelman
Meta-analysis can help identify sources of variation in the dose-response trend across studies. The hypothesis of no heterogeneity between studies beyond that explained by sampling variability alone can be tested by means of the multivariate extension of the Cochran Q-test (Ritz et al., 2008; Jackson et al., 2012). Formally, the null hypothesis is and the test statistic is defined as
The Fay–Herriot model for multiply imputed data with an application to regional wealth estimation in Germany
Published in Journal of Applied Statistics, 2022
Ann-Kristin Kreutzmann, Philipp Marek, Marina Runge, Nicola Salvati, Timo Schmid
One possibility to assess the quality of the model-based estimates is the comparison with the direct estimates. Brown et al. [6] propose a goodness-of-fit test for this assessment. The null hypothesis of the test assumes that the model-based estimates do not differ significantly from the direct estimates. The test statistic is defined as W is D degrees of freedom under the null hypothesis. The results of the test show that the null hypothesis, model-based estimates do not differ significantly from the direct estimates, cannot be rejected (see Table 5). According to [13], a useful diagnostic that measures the adequacy of the model is the correlation coefficient of the synthetic part of the FH-MI estimates and the direct estimates. For the federal states this correlation is 0.88 and for the planning regions 0.68.
Metabolic syndrome components and leukocyte telomere length in patients with major depressive disorder
Published in The World Journal of Biological Psychiatry, 2022
Yu-Chi Huang, Pao-Yen Lin, Yu Lee, Chun-Yi Lee, Yi-Ching Lo, Chi-Fa Hung, Cheng-Sheng Chen
For comparing continuous and categorical variables between patients with MDD and controls, the t-test and chi-square test were used, respectively. Spearman correlation was adopted to examine the potential correlation between categorical independent variables and continuous dependent variable. Linear regression was used to determine the most parsimonious model effective correlation of LTL between the MDD patients and control group, and then to examine the correlation of LTL between the MetS and its components in patient with MDD and the control group, respectively. In Model 1, demographic variables (sex, education, body mass index (BMI), cigarette use, marriage) and MetS were selected to control for; in Model 2, the aforementioned demographic variables selected in Model 1 and five MetS components were adopted to adjust for. Bonferroni correction was adopted for multiple comparisons. Moreover, linear regression was used for exploring the relationship between the group (MDD vs. controls) × HDL-C interaction effect and LTL by controlling for demographic variables; ANOVA and post-hoc analysis was used to compare the differences in LTL among subgroups. We log-transformed the relative LTL to normalise distribution, and the value of the log-transformation (LTL) was adopted in all analyses to represent the relative LTL. Unstandardised coefficients (B value) with 95% confidence intervals (CIs) were computed. All test-statistics were two-tailed, and differences were considered significant at p<.05.
A Prospective Randomized Controlled Trial: Alternative Approach to EEG Application to Reduce Electrode-induced Skin Injury among Ambulatory EEG Patients
Published in The Neurodiagnostic Journal, 2022
Sumika Ouchida, Armin Nikpour, Greg Fairbrother
Significant by-group differences were not noted for patient mood, patient comfort, or EEG reading quality (Table 3). Sleeping position (side, supine, or prone) was assessed against inflammation scoring for both groups, with no significant associations determined. The sleeping position was also assessed against comfort for both groups. No associations were located in the control group. However, in the intervention group, the supine sleep position was found to be significantly associated with “how strongly did the tactile sensations (itchiness or aching) attract your attention?” than for patients who slept prone or on their side (Kruskal–Wallis Test Statistic: 8.0; P = 0.02). A modest positive correlation (Spearman’s rho = 0.28; P = 0.01) was identified between total scores on patient mood and patient comfort, regardless of group. Patient mood total was not found to be associated with inflammation levels or EEG quality.