Sceintific production on chronic non-communicable diseases such as overweight, obesity and diabetes has grown in recent years.However, there is no comprehensive overview of the study designs and statistical analysis methodologies commonly used for risk assessment of research conducted on this topic.In order to evaluate risk assessment in this field, 1190 relevant documents downloaded using markers were included: ("Risk-Factors Associated with Overweight" or "Risk-Factors Associated with Obesity" or "Risk-Factors Associated with diabetes") and (microbiome or diabetes) and "multivariate analysis", in publications retrieved between 2017-2022 from the PubMed database; Specific parameters of title, journal, year of publication, authors, country of origin, institution, authors, keywords, among others, were analyzed.Data analysis was performed in three stages: 1. Descriptive; 2. Networking for Bibliographic Coupling Analysis (Network); 3. Strength of association of bibliographic coupling (Normalization).Data Visualization comprised: a.A mapping of the conceptual structure with Multiple Correspondence Analysis (MCA) for qualitative variables; b Network mapping.Grouping (Clustering) of K-means to identify groups of documents that express common concepts.To automate the data analysis and visualization stages, the open source tool (bibliometrix R-package), developed in R language, was used.68.6% of the variability of the ables.Most study methods -study design, analysis methodology -are based on continuous measurement variables, commonly using methods based on distributions such as normal for statistical analysis.In clinical epidemiology, where statistical methodology is applied to the study of diseases, their occurrence, distribution and relationship with explanatory variables, confronts us with variables of interest with distributions that are not continuous.Commonly continuous measurement variables become dichotomous