Dear all,
I'm having trouble with getting the historgram to give me 25 bins. I'm using
"break", but setting it to 25 or less seems to return too few and setting it to
26 or more seems to return too many. Any suggestions?
Thanks,
Phillip.
> temp <- hist(a,breaks =25)
> temp$counts
[1] 6 20 35 95 156 180 195 137 103 40 21 8 2 1 1
> temp <- hist(a,breaks =26)
> temp$counts
[1] 2 2 8 6 8 9 19 16 43 43 57 63 83 59 74 77 70 72 55 58 51 38 22 18 14
[26] 11 6 11 1 0 1 0 1 1 0 1
>
-------------------------------------------------
Phillip Y. Lipscy
Perkins Hall Room #129
35 Oxford Street
Cambridge, MA 02138
(617)493-4893
lipscy(a)fas.harvard.edu
Ph.D. Candidate
Harvard University, FAS, Department of Government
-------------------------------------------------
On Thu, 7 Nov 2002, Tao Li wrote:
[...]
> i discussed the regression coefficient intepretation for the standard case
> of continuous varible. but the incumbency paper is discrete variable. i
> could not answer this question well. could any of you send an email to the
> class list to explain more upon it? thanks.
>
> Tao Li
> --------------
> www.people.fas.harvard.edu/~li7
Hi everyone, when interpreting coefficients of regression equations,
whether x is discrete or continuous, you can use a sentence like this:
when x1 goes up by 1 unit, y goes up by b on average (or the estimate of
E(Y|X) goes up by b exactly), holding constant the other explanatory
variables, and given the veracity of the model assumptions.
Whether x is continuous or discrete doesn't really matter. It is still
going up by 1 unit. The key for you, however, is to make sure that one
unit makes sense. so in the incumbency advantage example, a 1 unit
increase means either -1 to 0 or 0 to 1, and nothing else. it also means
that the effect is the same for either change -- an assumption you're
making when running the model. a good thing to do would be to check
whether the assumption is right. How would you do that?
Gary
: Gary King, King(a)Harvard.Edu http://GKing.Harvard.Edu :
: Center for Basic Research Direct (617) 495-2027 :
: in the Social Sciences Assistant (617) 495-9271 :
: 34 Kirkland Street, Rm. 2 HU-MIT DC (617) 495-4734 :
: Harvard U, Cambridge, MA 02138 eFax (928) 832-7022 :
correct. as i said in class, there are 2 sections from 7-9. one for each
hour, obviously. i am going to discuss major problems in hw5, review
Gary's probability lecture(answer questions) and HAND OUT hw6. thanks.
On Thu, 7 Nov 2002 dhopkins(a)fas.harvard.edu wrote:
> Hey Tao,
>
> I just wanted to make sure I'm correct that we have section tonight from 7-8 PM
> or 8-9PM in M-16.
>
> Many thanks!
>
> Best,
> Dan
>
>
Dear Friends,
Dave suggested that I pass along some of his advice from a conversation we had
on Monday related to the readings for the class. Of the Netter assignment, he
highlighted the final three sections of chapter five as the most essential--
there, it is worth understanding exactly what Netter et. al. are doing and
why.
At the same time, while he said that some of the material in Chps. 6-11 won't
be directly tested or utilized in Gov1000, it is all material that political
scientists should be familiar with, since it is all material we might see in
practice. Dave recommends spending an hour per chapter: that should be
sufficient time to become familiar with the subjects, but not to follow every
line of every proof.
Best,
Dan
it probably makes more sense to change it to wednesday night 9-11. in this
way i can help answer homework questions. it will still be G-2 of littauer
basement. if no one objects, i'll start this wednesday. thanks.
the section will still be thursday night.
Tao Li
--------------
www.people.fas.harvard.edu/~li7