Question

What are the differences between results that demonstrate a correlation between two variables and results where a regression is run using two variables? 

Think about your future clinical role and provide a clinical example of variables that you may want a correlation analysis run and explain. 

Think about your future clinical role and provide a clinical example of variables that you may want a regression analysis run and explain.

Answer 1

1.

Results that demonstrate a correleation between two variables shows the linear relationship between those those two variables under study. Correleation measures the extend and direction of this linear relationship between the two variables. Here unlike regression, we do not define a dependent and independent variable & thus there is no difference between the dependent and independent variable i.e Corr(X,Y)=Corr(Y,X). Now Regression between two variables describes how a independent variable is numerically related to a dependent variable. Here we hypothesize a model of the linear relationship, and use the estimates of the parameter values to develop an estimated regression equation. Regression analysis is used for for prediction and forecasting. Given the numerical value of the independent variable, what would be the value of the dependent variable.

2.

In Clinical trials, suppose we have to want to test the relationship between Age and some disease, say Poliomyelitis occurence. This we are doing so we can focus our study for clinical trials on that particular age group. We randomly select a sample of 1000 people with Poliomyelitis disease and record their corresponding ages. Here finding the correleation would enable us to know whether the occurence of this particular disease increases with age, decreases with age of remains constant over age.

3.

In Clinical trials, suppose we want to know the effect of weight on blood pressure. Here weight is the independent variable, while blood pressure is the dependent variable. Here we want to know what would be the effect on blood pressure of different weighted individuals so we can decide the particular dosage for them (Higher BP patients would require a comparatively higher dosage of the drug).