Statistical analyses. Means are reported with standard deviations. For all statistical tests, normality was evaluated using the Kolmogorov-Smirnow test (when n > 50) and the Shapiro Wilk test (when n < 50). For normal distributions, we used a t-test (paired or unpaired, t). For non-normal distribution, non-parametric tests were used: (a) The Wilcoxon signed-rank test (Z) for within-subject comparisons; (b) the Wilcoxon Mann Whitney test (U) for two-sample comparisons. Finally, a Chi-square test (X2) was used to compare gender between participants having chosen French and those having chosen German to measure the independence of both variables in a 2-by-2table. Behavioral results were reported in the format: Test (degrees of freedom) = Statistic, p = value. All tests were two-tailed and a p-value < 0.05 was considered statistically significant. All statistical analyses on epidemiological and behavioral data were performed with SAS System V9.4 (SAS Institute, Cary, NC, US).
Statistical analyses. Means and medians of continuous periodontal variables and percentage of sites with PD≥5mm or CAL≥4mm were calculated for each subject and group. Considering the frequency distribution of the values above the threshold of detection, undetectable serum IL-6 values were recorded as 0.5 pg/ml. Natural logarithm transformation was applied to non-normally distributed variables. Continuous and dichotomous variables were compared among groups using a two-sample student t-test or χ2 test, respectively. Multivariate linear regression analysis determined predictors of serum IL-6 and CRP. Our transplant group sample size permits detection of an effect size of 0.18 at a p=0.05, in a multivariate analysis with 6 predictors and β≤0.2. Pearson correlation tested associations between serum IL-6, CRP and continuous periodontal variables. Partial correlations test was used to test the association between serum IL-6 and periodontal tissue IL-6 mRNA levels while controlling for diabetes. p≤0.05 was considered statistically significant. 5 iCycler iQ PCR detection system software, Bio-Rad.
Statistical analyses. Means and standard deviations were calculated for each of the TCI-R 7 scales and 29 subscales. Their internal consistencies were assessed according to the Cronbach's alpha coefficient. Gender differences in the mean scores of the TCI-R scales and subscales were explored using multivariate analysis of covariance in which age was set as the covariate variable in order to control for its influence. F statistics, p values and effect sizes (partial η2) were estimated. Linear associations among the 7 dimensions of the TCI-R were analyzed using a series of Pearson correlation coefficients.The factor structure of the TCI-R was analyzed through a Principal Component Analysis (PCA) with promax rotation. Temperament and character subscales were factor-analyzed separately because the relationships among the temperament and character dimensions are nonlinear and therefore cannot be adequately specified by the linear assumptions of factor analysis [5,35].Concurrent validity of the TCI-R was examined by calculating the Pearson correlations between the TCI-R dimensions and measures of the Big-Five personality dimensions (IPIP- 50), trait impulsiveness (BIS-11), depressive symptoms (BDI-II), suicidality (SBQ-R), and life satisfaction (SWLS). All statistical analyses were conducted using the SPSS version 19 (SPSS, Chicago, IL). The level of statistical significance was defined as P less than 0.05 (5%).
Examples of Statistical analyses in a sentence
Statistical analyses were conducted by FDA to determine whether the financial interests/arrangements had any impact on the clinical study outcome.
Statistical analyses must be conducted in accordance with international statistical reporting standards (Altman DG, Gore SM, Gardner MJ, Pocock SJ.
Statistical analyses for possible exposure-related effects12 on survival used Cox’s (1972) method for testing two groups for equality and Tarone’s (1975)13 life table test to identify exposure-related trends.
Diurnal fluctuations in brain volume: Statistical analyses of MRI from large populations.
Statistical analyses of the responses to LHRH injection, and melatonin and prolactin concen¬ trations during development consisted of one- and two-way analyses of variance.
More Definitions of Statistical analyses
Statistical analyses. Means and standard deviations (s.d.) were calculated for silage fermentation variables determined for samples prior to drying and are presented in Table 1. Table 1 Mean (s.d.) silage total solids concentration (TS) and fermentation variables (g/kg TS, unless indicated otherwise; except pH) prior to dryingSilage fermentation variables2Harvest1 TS/g kg-1 Note: 1 Harvest 1 = 12 May, Harvest 2 = 9 June, Harvest 3 = 7 July 20112 LA = lactic acid, AA = acetic acid, PA = propionic acid, BA = butyric acid, EtOH = ethanol, FP = total fermentation products (LA + AA + PA + BA + EtOH), LA/FP = lactic acid as a proportion of total fermentation products, NH3-N = ammonia-N Methane production was determined in duplicate batch digestion tests for each of the three replicate samples per treatment. These duplicate values were averaged to give a single value for each sample for subsequent statistical analysis. Herbage chemical composition and specific CH4 yield data were analysed as a split-plot design using the MIXED procedure of SAS, Version 9.1.2. The structure used harvest date as the main plot, with a 2 (ensiling; pre- or post-ensiling) × 2 (drying method; thermal or freeze drying) factorial arrangement of treatments within the sub-plot, and with the effect of replicate block being accounted for within the main plot. The Tukey adjustment for multiple comparisons was used in testing for differences between means.The CH4 production at each individual sampling day (2, 5, 8, 13, 19, 26 and 36) over the 36 day incubation period (i.e. CH4 production over time) was analysed as described above, but with the repeated measures effect of sampling day also being accounted for.
Statistical analyses. Means and standard deviations of all measured values were calculated for each condition. General Linear Model was applied to all set of data, followed by a multiple comparison, the Tukey’s test. Significance value was p<0.05. Minitab 17 (Minitab Inc., USA) statistical program was used for data analyses. Sensory scores were represented as the mean of each panelist’s scores. Means and standard deviations of the peak areas of volatile organic compounds were calculated for each condition. General Linear Model was applied to a set of VOC data, followed by a multiple comparison, the Tukey’s test. Significance value was p<0.05. The data of the peak areas of VOC and sensory scores of overall acceptability were submitted to Principal Components Analysis (PCA) for qualitative classification. For the quantitative estimation of the biochemical changes and the microbial population of total viable counts, Pseudomonas spp., LAB and sensory quality, PLS regression (PLS-R) models were calculated using the volatile compounds as input variables (predictors) and the quality parameters as output variables (response). The leave-one-out cross validation (LOOCV) technique was applied to evaluate the performance of the models. The Minitab 17 (Minitab Inc., USA) statistical program was used for all mathematical data treatments and statistical analyses.The PLS model is a bilinear regression model that extracts a small number of factors, which are a combination of the independent variables, and uses these factors as a regression generator for the dependent chemically and biologically measured variables(Maleki et al, 2006). PLS regression is known for its simplicity, robustness, predictability, precision, and clearly quantitative explanations. Despite this, PLS regression does not present a quantitative relationship between predictor variables and response variables, and it does not support re-use of model algorithms between different instrumentations (Li et al, 2012). Some validation parameters of model are evaluated such as X Variance, Error, R-Sq, PRESS and R-Sq (pred.). Prediction errors (PE) or standardized residuals for each of individual prediction point were calculated with the equation (3.5) below and it can be served as the overall performance of the model (Oscar, 2009). The percentage of PE above 70% present that the prediction model can be used for prediction of test data in the acceptable range (Oscar, 2005). %PE = (PEin/PEtotal) x 100 (3.5) Where, PEin : The number of PE...
Statistical analyses. Means (±standard deviations) of demographic and treatment variables were determined for the entire sample and for the three treatment subgroups. To verify similarities between treatment sub- groups before the initiation of treatment, Student's t-tests and a Chi- Square test (for gender) were performed. To evaluate treatment outcomes, paired Student's t-tests were performed for the entire sample and for the three treatment subgroups for the primary smoking outcome measures (cigarettes per day, FTND scores, and exhaled CO levels). Student's t-tests were then performed between the active treatment subgroups (PGC- and bupropion HCl-treated) and the inactive pill placebo subgroup for the primary smoking outcome measures.For the PET data, the variable studied was the smoking-induced percent change in 11C-raclopride binding potential (BPND), defined as 100 *( BPND before smoking − BPND after smoking)/BPND before smoking, for thevolume-corrected mean (of the left and right VCD/NAc) BPND value. Asingle BPND measure was used here, based on our prior report demonstrating nearly identical BPND changes with smoking for the left and right VCD/NAc (Brody et al., 2004).For the central study analysis, a repeated-measures ANOVA was performed, with smoking-induced BPND percent change before and after treatment as the repeated measure and treatment group as the between- subject factor. This same analysis was also performed with pre-smoking BPND as the repeated measure, as a marker of change in baseline intrasynaptic DA from before to after treatment. Pearson Product Moment Correlation Coefficients were calculated to examine relation- ships between pre- to post-treatment changes in smoking-induced BPND reductions and changes in variables that might be expected to be associated with smoking-induced DA concentration change, namely total puff volume, quit status, and craving alleviation. Statistical tests were performed with SPSS version 16.0 (SPSS; Chicago, IL).
Statistical analyses. Means and 95% confidence intervals were obtained from a general linear model for longitudinal measurements with a covariance matrix taking into account the presence of multiple scar sites within a patient and the repeated measures over time. A random patient effect was used to model the correlation between scar sites from the same patient. For the repeated measures over time an unstructured covariance matrix was used. If the distribution of the model residuals was right- skewed, the outcome was log-transformed (natural logarithm), but figures were created after back-transforming to the original scale (in which case they refer to geometric means and their 95% CI).Two different models were used. In the first model the evolution over time was evaluated on all repeated measures. In a second model the post-baseline values were compared between the two groups. The analysis was restricted to post-baseline values and the baseline value and the age of the scar were added as covariates in the model, their effects being allowed to vary over time (by including interactions with time). Hence, the plot depicts the mean value of scar sites of mean age and mean baseline value. The scar age (time in months (mo) between wound closure and baseline assessment) was log-transformed (natural logarithm) to downplay the potential influence of extreme observations.All analyses were performed using SAS software, version 9.2 of the SAS System for Windows.
Statistical analyses. Means, standard deviations, percentages, and correlation coefficients were calculated. Five hierarchical regression analyses were conducted where Facebook addiction, stress, general health, sleep quality, and quality of life were dependent variables. Independent variables introduced in subsequent steps can be found in Tables 2 and 3. For all linear regression analyses, preliminary analyses were conducted to ensure no violation of the assumptions of normality, linearity, multicollinearity, and homoscedasticity. All tests were two-tailed and the significance level was set to α = .05.
Statistical analyses. Means were compared by Fully Factorial MANOVA with diet treat- ment as the factor and data on individual larval weight as the depen- dent variable. Statistical significance was determined by the Tukey HSD multiple comparison test (Tukey, 1951). For insect bioassays with multiple doses, POLO-PC was used for probit analyses (Robertson et al.,1980). The Fisher exact test was used to test for significant differences among mortalities (n ¼ 16).
Statistical analyses. Means and standard deviations across outcome measures are presented in Table 3. A series of Generalised Linear Mixed Models (GLMMs) using SPSS (Version 24) GENLINMIXED procedure assessed intervention effects across outcome measures. Each GLMM included two nominal random effects (participant, dyad) and one ordinal fixed effect (time: pre, post, follow-up). The traditional ANOVA repeated measures model requires the following assumptions to be satisfied: normality, sphericity, and independence of observations. The GLMM “robust statistics” option accommodates violations of normality. Violations of sphericity was accommodated by changing the covariance matrix from the default of compound symmetry to autoregressive. Finally, by specifying the multilevel nature of the current data (participant nested within dyad) in the GLMM syntax, GLMM accommodated intra-dyad dependencies in the outcome measures.Compared to the traditional ANOVA repeated measures model, GLMM is less sensitive to participant attrition because it does not rely on participants providing data at every assessment point; the GLMM maximum likelihood procedure is a full information estimation procedure that uses all the data present at each assessment point. This reduces sampling bias and the need to replace missing data. GLMM is able to use the data present at each assessment point, this is because time (pre, post) is interpreted as a Level 1 variable that is nested within participant at Level 2, which is itself nested within dyad at Level 3.To address possible inflation of familywise error rate, outcomes were evaluated at Bonferroni corrected levels, whereby alpha was divided by the number of subscales within each measure.