couldn’t agree more, slouiscar…all facets of LOSs are fair game, but c’mon, devoting as much of the exam as they did to very, what i would call, minor points doesn’t seem like a fair way to separate the better prepared candidates from the less prepared candidates.
“couldn’t agree more, slouiscar…all facets of LOSs are fair game, but c’mon, devoting as much of the exam as they did to very, what i would call, minor points doesn’t seem like a fair way to separate the better prepared candidates from the less prepared candidates.” This was exactly how many of us felt last year.
i’m with mark, need the intercept
I now agree with Mark, for the same reasons why a 1% change in sales will not have a 1% change in EBIT (DOL) Wish i had think kind of critical thinking on the exam
got it. If that was the q I agree you go +1. I do not recall wording or what I answered. If a question asks what is the change in exchange rate and that is = Y, well there you go, you need to calc the value of Y using the model. If you ask a q and include a phrase like holding all else constant… that phrase is taken right from the multi reg section on interpreting the impact of indiv indep variables. I don’t remember the q. The concept itself is so ridiculously easy it is quite painful to acknowledge that I don’t really remember how I interpreted it at the time.
I am going to puke if I missed this question. I can do time series with no problem and I might have missed this…
were there not 2 quant ones, one asked what would be the change of 1 variable and it was just that B1 x the 1/2 or whatever it was? but then the one that was the .8 and do you or don’t you add intercept, i think it asked what would be the change of the dependent so i added the intercept. hope that was right.
You shouldn’t use the intercept, they were looking for the incremental change.
2 tailed test, correlation .339 and PPP was right for me. Banni how did you do overall? I will on the border…Anything can happen.
The dependent is defined as change of the exchange rate. The problem didn’t ask you to figure out the change in the change of the exchange rate. So you should add intercept.
I thought it said how would it affect the value, not what was the value.
I believe it said to just hold the third term constant, not all else constant. I included the intercept.
intercept definitely included…it is a part of the model…to not include it in determining effects on the dependent variable would be a model misspecification.
Yep, I definitely ran the whole model.
FACK was the model predicting change in value? FACCCCCKKKKKKKKKKKKKKKKKK
i dont think we had to run the whole model… it said holding one constant what whould the change be for a 1 unit change in the other one… it not ask us to predict the value…
agreed … if you had y = 2x + 3z and asked what the effect was by changing x by one and holding z constant, the answer is that y would increase by 2. This is linear regression, so you can think of the coefficients as sensitivities to changes in the independent variables, and if you calculated the total derivative, dy = 2 dx + 3 dz. x and z are independent. So dy/dx = 2. To calculate it out in full, take z = sqrt(2) and x = 1.5. Then y = 2*1.5+ 3 * sqrt(2) = 3 + 3 sqrt(2). Now change x by 1, so y is now 2*2.5 + 3*sqrt(2) = 5 + 3*sqrt(2). y changed by 2.
I ran the whole model - something like 5.2 - the reason being, if you want to know the change in the dependent (even with all else constant) you have to add the intercept and b2
it was 0.85 and you shouldn’t add intercept. what they asked in the question was how dependent variable would change if we change one independent by 0,5 HOLDING OTHER VARIABLES CONSTANT. corect answer is b1*0,5. that is what definition of slope coefficient in linear regression model suggests. if we follow argumentation that we should add intercept what would be the value of second coefficient b2? i don’t think that we can assume that it is 0.
rado.pl Wrote: ------------------------------------------------------- > it was 0.85 and you shouldn’t add intercept. what > they asked in the question was how dependent > variable would change if we change one independent > by 0,5 HOLDING OTHER VARIABLES CONSTANT. > > corect answer is b1*0,5. that is what definition > of slope coefficient in linear regression model > suggests. > > if we follow argumentation that we should add > intercept what would be the value of second > coefficient b2? i don’t think that we can assume > that it is 0. compl AGREE. it was 0.8 - I need this for 6/6 on Quants