Interpretation of Chi-square test in pediatric investigation

Abstract

Qualitative variables are often used in pediatric research, for example, newborn gender (male, female) or malnutrition level (lesser, moderate, severe). In order to determine the association or independence of two qualitative variables with a certain level of significance, a frequently used statistical tool is available: chi–square test (X²). The present article explains the theoretical background of the test, the methodology employed for calculating the X² statistic and its correct interpretation, exemplifying these concepts using a real investigation. In simple terms, chi–square test (X²) compares the results observed in an investigation with a set of expected results calculated under the assumption that both variables are independent from each other. The difference between the observed and expected results is summarized in the value of the X² statistic, which has an associated p–value; according to its magnitude, the independence hypothesis is either supported or rejected. In consequence, the application of the chi–square test (X²) allows the investigator to determine if two variables are associated or are mutually independent, an affirmation that can be statistically sustained.