Dummy variables, predict stock market prices. PLEASE HELP!

[janegoodallschimp]

Hi all. I'm desperate. I have been sitting at my desk for 11 hours doing math and this is the last one. We have an exam in a week and I just don't understand this concept.

"Some say there's a January effect in the stock market (that the stock market rises extra much that month). You have data that relates to the return on the stock market over several months time. On average the return has been 2.9% in January, and 3.1% all the other months. What values does a and b get?"
Return^ = a + b*January
Dummy variables: January (value = 1), rest of the months = 0.

What is the value of a and b?

Here's an earlier example in the same style:
rent^ = 2115 + 713 * two bedroom apartment (which becomes a dummy = value 0) = 2115

 

[Dr.Peterson]

I'm not familiar with this style of problem, but it looks to me as if you want to find a and b such that

a + b(1) = 2.9%​

a + b(0) = 3.1%​

Can you do that?

 

[JeffM]

I'm not familiar with this style of problem, but it looks to me as if you want to find a and b such that

a + b(1) = 2.9%​

a + b(0) = 3.1%​

Can you do that?

Dr. Peterson is correct.

A dummy variable is an on\off switch. It really makes no difference whether you pick January as on or off. Nor does it make any difference whether you pick 0 or 1 for on although it's conventional to pick 1 for on.

Why just two values?

It's either January or it isn't.

The whole idea is to determine whether the coefficient of the dummy variable is statistically significant within a specific model. If it is not, then you conclude that adding that variable to that specific model is not helpful. The dummy variable might turn out to be statistically significant in a different model.