[Research protocol VII. Validity and reliability of the measurements]

The concept of validity in research refers to what is true or what is close to the truth. It is considered that the results of an investigation will be valid when the study is free of errors. The errors or biases appear in the development of research, are due to methodological problems and, in general, can be grouped into three: selection bias, measurement bias and confusion bias. In this article, measurement biases will be addressed; this type of error has three axes: 1) the research subject, 2) the instrument for the measurement of the variable (s), and 3) those who make the evaluation of the measurement (s). To improve the obtaining of data and to prevent errors, some strategies can be followed: in every study protocol, it is necessary to include the operational definition of the variables; the subjects that will carry out the measurements or surveys must be trained. If measuring instruments are used, their proper functioning must be verified; when questionnaires are included, they must be validated in the language in which they will be applied, they have had a process of adaptation to the language of the participants in the study, and self-applicable ones are preferred. It is necessary to quantify the variability of the measurements from the statistical point of view to increase the validity of a study.