Question

True or false. effect modification and correlation value

1) True or False? An initial analysis finds that the estimated slope for the regression line predicting SBP as a function of “Fizzy Delicious Lite” soda consumption (in liters) is b1 = 8.02. Fizzy Delicious Lite soda consumption is associated with Yummy Yummy candy consumption, and Yummy Yummy candy consumption is associated with SBP. When Yummy Yummy candy consumption is added to the model, the estimated effect of Fizzy Delicious Lite consumption on SBP is b1 = 0.96. This is an example of effect modification.

2) True or False? The value for correlation is the same no matter which variable is the response and which variable is the explanatory, but regression lines will be different depending on which is variable is response and which is explanatory.

Answer 1

(1) It is TRUE because the association or slope coefficient between SBP as a function of Fizzy is 8.02 and when third variable Yummy Yummy candy is added, the slope changed to 0.96. So, this is perfect example of effect modification where association between two variable is changed due to the introduction of third variable

(2) It it TRUE because the correlation is independent of which variable is response and which variable is explanatory, correlation only depends on array 1 and array 2, but not on variable as response or explanatory. Regression line is specific to response variable and predictor variable, so we can change response and variable in regression line otherwise we will get a new regression line.