Sample Questions for SAS Visual Statistics Certification

[xliu1]

I am reviewing the practice questions for SAS Interactive Model Building using SAS Visual Statistics 8.4 on SAS Viya, and I am quite confused by correct answers for these two questions:

Q1:

In my opinion, B and D are not linear while A and C are linear

Q2:

I think D should be the correct one.

Just want to post here to get confirmation on these very basic concepts. Or am I not reading these questions correctly?

[MarkusWeick · Accepted solution]

Hi @xliu1

concering Q1: I think you are rigth. The X as the Variables are linear in answer A. and C. 

Concerning Q2: I think you are wrong.

The expession of the linear regression would be .... - 0.3696 x RunPulse + ... .

So a one unit decrease in RunPulse would lead to a 0.3696 increase in the response and not vice versa.

Hope, this helps.

Best, Markus

[xliu1]

@MarkusWeick Thank you for your response. Now I can see the wording is wrong for option D. But their correct answer C is not right. For one-unit increase in RunTime, should there be expected 2.6287 decrease in response as the equation follows102.9345+...-0.36963*RunPulse-2.62865*RunTime...? 

[MarkusWeick]

@xliu1: I agree with you. It seems to be a typo.

[Mark2010]

OK, so the run pulse increase/decrease is a typo and will be changed to decrease. My apologies for this.

But on the linear model question, the answer C is correct. I had a consult with one of our stats instructors, and here is the scoop:

The linearity of the model is based on the structure of the parameters (the b's) and not the predictors (the X's). The X1X2 and X3 (sorry, I can't do subscripts and superscripts here) will give us a linear model with non-linear effects. The division element in answer choice C (b2/b1) is the part that causes the non-linearity in the model. 

[MarkusWeick]

@Mark2010: Thanks. Now I see, that you are right. I consulted one of my text books.

@xliu1: Linear model - Wikipedia is a good short presentation, which shows different uses of linear in different statistical contexts.

[xliu1]

Thanks all for quick responses and sharing the link. It clears up my confusion. Very much appreciated!