Author 
Message 
Senior Manager
Joined: 14 Jun 2007
Posts: 397

looking for some reassurance [#permalink]
Show Tags
14 Sep 2008, 16:16
I'm thinking about switching my MS from one in marketing to an MS in Statistics. There was some overlap between the two, and I just figure that taking the extra (tougher) courses in Stats and getting the MS in Stats is a better investment... I'd still do an independent study or two with a marketing professor to get some research done in the area I am interested in before I graduate.
I know this is the right decision, but I am nervous because the chance of me getting a 4.0 in Stat is much slimmer than in Marketing. I just need reassurance that adcoms would be more impressed with a student with a 3.7 in Stat vs. 4.0 in marketing.



Manager
Joined: 24 Feb 2007
Posts: 56

Re: looking for some reassurance [#permalink]
Show Tags
16 Sep 2008, 16:17
Just speaking as somebody inside looking back, I don't think a 3.7 v 4.0 will make or break your application. Your study with the marketing professor, and what you have to show for it via your SOP, will actually contribute more to your overall application package. Grades don't impress them (much), since they get straight4.0s every other applicant. Somebody who shows logical thought and research potential is a much more attractive applicant.
That being said, having taken a few weeks of stats course, I can tell you a good foundation in statistics will be nice. Our class is a combination of OB, Marketing, and CIS types, and the prof took one week to "review" confidence intervals, after which everybody sort of walked out of class looking very much "THAT was just a review?!". If you're curious, the text we will be using is Applied Linear Statistical Models, by Kutner, Nachtsheim, Neter, & Li, supposedly a popular book amongst PhD stats courses. Of course, once again, each program differs, so what I'm saying may not apply to other programs.



Senior Manager
Joined: 14 Jun 2007
Posts: 397

Re: looking for some reassurance [#permalink]
Show Tags
28 Sep 2008, 11:24
Kutner is definitely not an easy read, and I feel the notation used is not precise. Capital letters are generally reserved for random variables, yet my whole understanding of OLS is that our regressors are under experimental control.
Also, where the book fails is when it uses calculus to find derivatives of what amounts to be functions of constants and not variables.



Manager
Joined: 25 Apr 2005
Posts: 196

Re: looking for some reassurance [#permalink]
Show Tags
29 Sep 2008, 21:22
anonymousegmat wrote: Kutner is definitely not an easy read, and I feel the notation used is not precise. Capital letters are generally reserved for random variables, yet my whole understanding of OLS is that our regressors are under experimental control. That probably depends on how you learned that stuff. In linear algebra (which is what almost all of econometrics is derived from), capital letters represent matrices while lowercase letters are for vectors. Hence the regression Y = XB + E (B and E should be beta and epsilon greek letters here of course) refers to the 'true' model which presumably applies to the whole population. In other settings you may have the distinction you brought up, but all of my econometrics books (Greene, Woolridge, Hayashi, Goldberger, Kennedy) use this notation.



Manager
Joined: 24 Feb 2007
Posts: 56

Re: looking for some reassurance [#permalink]
Show Tags
30 Sep 2008, 19:04
What cabro57 is saying is how I learned it... then again, my college stats class was in the economics department.
Kutner isn't a bad book... it just tries to do too many things at once, resulting in 1) Sometimes not very detailed and 2) Weighing a metric ton... even our prof just decided to toss the book when some of us started showing up with rollaboards, lol.



Senior Manager
Joined: 14 Jun 2007
Posts: 397

Re: looking for some reassurance [#permalink]
Show Tags
11 Oct 2008, 15:53
cabro57 wrote: anonymousegmat wrote: Kutner is definitely not an easy read, and I feel the notation used is not precise. Capital letters are generally reserved for random variables, yet my whole understanding of OLS is that our regressors are under experimental control. That probably depends on how you learned that stuff. In linear algebra (which is what almost all of econometrics is derived from), capital letters represent matrices while lowercase letters are for vectors. Hence the regression Y = XB + E (B and E should be beta and epsilon greek letters here of course) refers to the 'true' model which presumably applies to the whole population. In other settings you may have the distinction you brought up, but all of my econometrics books (Greene, Woolridge, Hayashi, Goldberger, Kennedy) use this notation. When my text uses matrix methods it uses boldface notation... so I think the authors are just taking shortcuts and not being careful with notation visavis what is random and what is not. I guess it is just implied that you are supposed to know when gears shift between random and not, but I think it is bogus. Likewise, I think it is lame to see a supposedly revered text find the derivative of a function with respect to Beta1 in order to minimize squared errors, when Beta1 is a constant, so it doesn't make sense.... if they were precise they would introduce a variable, theta. But I guess they didn't want to waste ink and add any more bulk to the text... (sarcasm).




Re: looking for some reassurance
[#permalink]
11 Oct 2008, 15:53







