China
Africa
Explore chapters and articles related to this topic
How Will the Data be Analyzed?
Published in Trena M. Paulus, Alyssa Friend Wise, Looking for Insight, Transformation, and Learning in Online Talk, 2019
Once the overall need for multilevel modeling (and the levels at which groups need to be modeled) has been established, the model(s) can be specified. In doing so the researcher can look for both fixed effects (which represent relationships between variables across the whole sample of data) and random effects (which represent the effects of group membership on the overall level of a variable and/or relationships between them) (Garson, 2013). In addition, researchers must decide whether variables should be analyzed as univariate measures (considered separately from each other) or as a multivariate one (considered together). Proper set-up of the model specification is important to ensure model convergence. Researchers often are encouraged to run several iterations of models before setting up the final model specifications to ensure the best model fit has been obtained (Heck, Thomas, & Tabata, 2013). For example, De Wever, Van Keer, Schellens, and Valcke (2007) explored, in iterative cycles, three different models to ascertain the best fit to explain the impact of role assignment in online discussion groups on knowledge construction, conducting post-hoc analyses to isolate the effect of variables such as task complexity on their outcome of interest.
Psychiatric Research
Published in M. Venkataswamy Reddy, Statistical Methods in Psychiatry Research and SPSS, 2019
These methods portray accurately the characteristics such as the type of frequency distribution of observed values and summarization figures of the characteristics of the univariate data. Thus, it helps in gaining familiarity with the situations and phenomena.
Multivariate Statistical Analysis—An Overview
Published in K. V. S. Sarma, R. Vishnu Vardhan, Multivariate Statistics Made Simple, 2018
K. V. S. Sarma, R. Vishnu Vardhan
Analysis of variables one at a time (each one separately) is known as univariate analysis in which data is described in terms of mean, mode, median, standard deviation and also by way of charts. Inferences are also drawn separately for each variable, such as comparison of mean values of each variable across two or more groups of cases.
Clinical characteristics, risk factors, and prognostic analyses of coronary small vessel disease: a retrospective cohort study of 986 patients
Published in Postgraduate Medicine, 2023
Yue Chen, Xiao Cui, Liujun Jiang, Xiaolei Xu, Chaoyang Huang, Qiwen Wang
Besides, the results showed that CSVD (OR = 2.490, 95% CI: 1.396–4.441, P = 0.002) was significantly associated with MACE events in univariate analysis but was not found to be significant in the result of multivariate analysis. In actual analysis, the results of univariate analysis and multivariate analysis are often different because the result of multivariate regression analysis is the effect of independent variables on dependent variables after excluding other interference factors (mainly confounding factors, but not intermediary variables). In our multivariate Cox regression model, confounding factors, such as age, diabetes, hypertension, and so on, were corrected, leading to no significant difference in P value of CSVD. However, it did not contradict with our conclusion that incidence of adverse cardiovascular events and revascularization in CSVD were higher than in non-CSVD. We expect to investigate the question by expanding the sample size and conducting RCTs in the future.
The impact of serum albumin-to-alkaline phosphatase ratio in cervical cancer patients treated with definitive chemoradiotherapy
Published in Journal of Obstetrics and Gynaecology, 2022
Cem Onal, Melis Gultekin, Guler Yavas, Ezgi Oymak, Sezin Yuce Sari, Ozan Cem Guler, Ecem Yigit, Ferah Yildiz
All statistical analyses relied on standard software (SPSS version 22; SPSS Inc. (IBM), Chicago, IL). A descriptive analysis was performed by calculating the mean, standard deviation, range and median. The primary endpoints of the study were overall survival (OS) and progression-free survival (PFS). A receiver operating characteristic curve (ROC) analysis was generated for the serum albumin levels, ALP levels and AAPR values to determine the cut-off values for predicting a recurrence and survival that yielded the optimal sensitivity and specificity. The Chi-squared (χ2) test or Student’s t-test was used to analyse the differences in the clinical and histological factors between patients with high and low AAPR levels. Time-to-death or progression was calculated as the period from the date of diagnosis to the date of death or first clinical or imaging evidence of disease recurrence. Both OS and PFS rates were estimated using the Kaplan–Meier method. The Chi-squared test or the Student t-test was used for univariate analysis. The multivariate analysis was performed using the Cox proportional hazards model, as defined by the hazard ratio (HR) and 95% confidence interval (95% CI), using all factors that were found to be significant by the univariate analysis. p < .05 was considered statistically significant.
A multicenter retrospective study on evaluation of predicative factors for positive biopsy of prostate cancer in real-world setting
Published in Current Medical Research and Opinion, 2021
Ben Xu, Gonghui Li, Chuize Kong, Ming Chen, Bin Hu, Qing Jiang, Ningchen Li, Liqun Zhou
All variables were analyzed descriptively with appropriate statistical methods: categorical variables were reported as frequencies (absolute and relative frequencies) and continuous variables by simple statistics (i.e. mean, SD, minimum, median, quartiles and maximum). Univariate analysis was carried out using various analytical methods based on the nature of variable. For measurement variable such as fPSA, PSAD, testosterone level, time interval of biopsy, t-test (normal distribution data) and Wilcoxon rank sum test (skewed distribution data) was used for the data meeting and skewed distribution respectively. The factors with p < .1 in univariate analysis were included in multivariate logistic regression model. Receiver operating characteristic (ROC) curves were used to assess the diagnostic accuracy of the predictors of positive biopsy. All statistical analyses were performed using the SAS 9.4 software (SAS Institute Inc., Cary, NC, USA). A two-sided p < .05 was considered statistically significant.