After a few biostatistics classes, I began fitting my first logistic regression model using my physician friend’s data on tumors excised from skin cancer patients. I realized that although we were very clear about the dependent variable we were trying to predict – a certain feature of the tumor – I really did not know how to pick the independent…

The importance of completeness of linear regressions is an often-discussed issue. By leaving out relevant variables the coefficients might be inconsistent.

But why on earth?! 

Assuming a linear complete model of the form:

z = a + bx + cy + ε.

Where z is supposed to be dependent, x and y are independent and ε is the error term.

Now we drop y to check…

Hi All;

suppose I am working on Job Satisfaction. I have a question Hours worked each week ________

 and the possible values are

(1) Extremely Dissatisfied 

(2) Dissatisfied

(3)Somewhat satisfied

(4) Satisfied

(5)  Extremely satisfied

Should I treat this variable ordinal or scale?