التحليل الاحصائي لمتغير الاستجابة الرتبي : دراسة مقارنة

الملخص EN

Response variables in biological phenomena vary between three types : numerical response variables, ordinal categorical response variables, and nominal categorical response variables.

In statistical studies, handling ordinal variables varies in accordance with the perspective of the statistical approach to the response variable.

Ordinal variables can be adopted as nominal categorical variables, which neglect the ordinal property of the categories.

Ordinal variables can also be treated.

as an ordinal categorical variable (discrete variable), in which case the ranking information can be utilized in establishing the predicted models.

In this study, the most important statistical methods that can be used to analyze data with an ordinal response variable have been investigated.

Among these methods are the Multiple Regression Method, and The Ordinal Logistic Regression Method.

The mechanism of building models and parameter estimations were theoretically exhibited, as well as reading the statistical significance of the regression coefficients in all the models in the study.

The application was carried out on a real sample of patients with osteoporosis.

Where multiple models were built to determine the most important factors affecting the likelihood of developing the disease.

The best model was diagnosed according to the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC).

The results of the statistical analysis demonstrated the superiority of the ordinal logistic regression model over the multiple linear regression model in its explanation of the relationship between the response variable and the covariates.