Sensitivity in Simple Linear Regression

[matt23]

Hi,

I was asked to do sensitivity analysis for my simple linear regression models but I do not know how.

If I have my simple PROC REG code

PROC REG data=dataname;
model Y=X;
run;

and I get my results. Is there anything I need to look at specifically for sensitivity or is there another procedure that would give me the sensitivity?

[matt23]

Would sensitivity just mean the coefficient for the variable ?

[PaigeMiller]

---

@matt23 wrote:

Would sensitivity just mean the coefficient for the variable ?

---

That's a reasonable understanding of the meaning of the word "sensitivity", but there are other reasonable meanings, and you need to clarify what is meant by "sensitivity" before actually doing the analysis.

[Reeza]

Sensitivity and specificity are statistical terms but would not apply to a linear regression with the standard model. Logistic regression would include a sensitivity analysis.

A more detailed description of your problem is required.

---

@matt23 wrote:

Would sensitivity just mean the coefficient for the variable ?

---

Likely not.

Statistical definition

https://en.wikipedia.org/wiki/Sensitivity_and_specificity

That being said, sensitivity analysis is a different topic which could possibly be what you're looking for here?

[matt23]

I'm not trying to argue as I have no idea but I found this on wikipedia:

Regression analysis, in the context of sensitivity analysis, involves fitting a linear regression to the model response and using standardized regression coefficients as direct measures of sensitivity. 

[PaigeMiller]

Someone asked you to do a sensitivity analysis. Until we know what that meant (and you know what that meant), there is no possibility of providing you an answer. Go back to the person who asked, and try to find out what was meant.