Salvavidas Estadístico

Determine the most appropriate statistical test for analyzing your results from a biomedical research or another field of science by answering three simple questions. This app can be used in any statistical analysis situation including two variables (bivariate statistics). It is especially useful in the field of research as well as any branch of science where this type of statistical analysis is required.

There is a very common difficulty among health professionals when analyzing the results obtained in a research procedure. The development of clinical activity and its preparation often makes it difficult to learn how to face methodological and statistical problems. At a teaching level, it is complex to develop the ability to perform mathematical logic analysis to face statistical problems, which requires time that is not compatible with clinical practice.

There are many manuals and books dedicated to this issue, as well as a great amount of mathematical software that allows the most complex operations to be carried out and high quantities of data to be handled. However, between the basic knowledge needed to analyze the data obtained in some experiments and the knowledge about advanced statistics and data handling, there is a gap that professionals often find insurmountable. Interacting in the world of information and interpreting it correctly is a basic process that every researcher must control in order to obtain correct conclusions that allow science to advance.

Thus, a method that facilitates the choice of the most appropriate statistical test for each situation is presented, allowing the researcherers to analyze their results without the need of having advanced mathematical or statistical training.

The objective of the physical device "Salvavidas Estadístico" and its app, is to provide a 3-step guide to carry out a bivariate analysis independently, applying the statistical concepts required in the different fields of science. Through the choice of the confrontation of two variables and the evaluation of their characteristics through 3 questions, the test that best suits each situation within a research study will be suggested, specifically in the analysis of results. Three questions are presented:

1: What kind of variables will be tested?

The first step consists of answering what type of confrontation is faced from two variables that can be quantitative or qualitative (dichotomous or polytomous). The options are:

- Quantitative vs. quantitative.

- Quantitative vs. qualitative dichotomous.

- Quantitative versus polytomical qualitative.

- Dichotomous qualitative versus dichotomous qualitative.

- Polythomous qualitative versus dichotomous qualitative or polythomous.

2: Is there a good fit to the normal distribution?

To analyze the normality of a quantitative variable, the Shapiro-Wilk test if n<30 or the Kolmogorov-Smirnov test if n>30 should be used. Both the Shapiro-Wilk test (for <30 cases) and the Kolmogorov-Smirnov test (for >30 cases) are interpreted in the same way. If the result is p>0.05 we can assume normality (parametric statistics) but if p<0.05 we will assume a bad fit with the normal distribution (non-parametric statistics). It will be necessary to answer Yes or No.

3: Are the variables paired?

Two variables are dependent, paired or matched only when they are directly related to each other (same individuals or cases at different times). Therefore, it will be necessary to evaluate whether the comparison to be established between the groups is paired or not. It will be necessary to answer Yes or No.

After answering these three questions, information will be given on the most appropriate test for the specific situation described.

Salvavidas Estadístico® is a registered trademark. 2020.