Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Testing joint hypothesis on means, ANOVA


From   Jules Perng <[email protected]>
To   [email protected]
Subject   Re: st: Testing joint hypothesis on means, ANOVA
Date   Tue, 7 Jan 2014 22:06:21 +0800

Thank you so much for your response!

I think this is what I'm looking for. Right now I am running:
regress Amount i.Factor_1 i.Factor_2

contrast g.Factor_1 g.Factor_2, effects overall

after which I believe I will use the results at the beginning which
look like this:

             |         df           F        P>F
-------------+----------------------------------
      Factor_1 |
(1 vs mean)  |          1        0.23     0.6303
(2 vs mean)  |          1        2.62     0.1057
(3 vs mean)  |          1        0.07     0.7958
(4 vs mean)  |          1        0.17     0.6846
(5 vs mean)  |          1        0.15     0.6952
      Joint  |          4        0.75     0.5564

If I could trouble you again, that would be great. I don't think this
takes into account the way I shaped my data, in which I have repeated
measures. So I wonder if there is a way to do so? I can't run
contrasts after doing ANOVA.

 My problem might be a bit abstract, so if it helps, one of the
factors is motives, and the other one is the giving under which type
of game.

My end result will look like:

                                Game 1         Game 2        Game 3
       F Stat
Everyone (total)     mean (s.d.)   mean (s.d.)   mean (s.d.)      -
Chosen b/c friend mean (s.d.)   mean (s.d.)   mean (s.d.)      0.89 (0.05)
Chosen b/c needy mean (s.d.)   mean (s.d.)   mean (s.d.)      0.73 (0.00)
Chosen b/c sick    mean (s.d.)   mean (s.d.)   mean (s.d.)      0.28 (0.10)

So I want the F-statistics on the right to be from a test that each
row below Everyone (total), as in each of the means in each of the
games (from that row), is significantly different from the first row.

(So what I ran was the following:
regress motive_amount i.why i.motive

 contrast g.why g.motive, effects overall)

I've been scratching my head on this for a few days, so if anyone had
an idea, that would be great!

On Sun, Jan 5, 2014 at 12:17 AM, William Buchanan
<[email protected]> wrote:
> Try -help contrast-.  You should also provide a bit more information regarding how you specify the ANOVA itself since that could also create issues when doing the other comparisons.
>
> Sent from my iPhone
>
>> On Jan 4, 2014, at 9:39, Jules Perng <[email protected]> wrote:
>>
>> Greetings all,
>> I have two factors and a dataset with repeated measures. I have
>> reshaped it so that it looks a bit like the following:
>>
>> Subject   Factor_1  Factor_2  Amount
>> 1             1                1               5
>> 1             1                2               7
>> 1             1                3               3
>> 2             3                1               5
>> 2             3                2               2
>> 2             3                3               8
>> 3             2                1               4
>> 3             2                2               8
>> 3             2                3               4
>> 4             5                1               4
>> 4             5                2               3
>> 4             5                3               7
>> 5             3                1               9
>> 5             3                2               5
>> 5             3                3               8
>> 6             4                1               8
>> 6             4                2               9
>> 6             4                3               4
>>
>> Where each subject will select one from five choices of Factor_1 and
>> select an Amount for each of Factor_2.
>>
>> I will be arranging the data like so:
>>
>>                                  Factor_2: 1   Factor_2: 2  Factor_2:
>> 3     F stat.
>> All Factor_1 (total)
>> Factor_1: 1
>> Factor_1: 2
>> Factor_1: 3
>> Factor_1: 4
>> Factor_1: 5
>>
>> Where I will have the means for each within the matrix.
>>
>> I would like to run joint hypothesis tests stating that all of the the
>> row means for each of 1 through 5 of Factor_1 are jointly different
>> from the total means (in the first row). I would find an F-stat for
>> each, so there would be 5 F-statistics (the first row would be blank).
>>
>> I have been quite stuck on how to do this. I found this reference:
>> http://www.stata.com/support/faqs/statistics/multiple-comparison-tests/
>> - and it seems that I should run anova, and then test. However, the
>> example doesn't show how to compare one level to all of the levels
>> together. I also played a bit with the manova command, but it didn't
>> give me what I wanted.
>>
>> It would be great if someone knew the response to this. I have combed
>> the internet so hopefully this is not a repeated inquiry.
>>
>> Thanks so much.
>>
>> --
>> Jules Perng
>> 503-943-9574
>> *
>> *   For searches and help try:
>> *   http://www.stata.com/help.cgi?search
>> *   http://www.stata.com/support/faqs/resources/statalist-faq/
>> *   http://www.ats.ucla.edu/stat/stata/
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/



-- 
Jules Perng
503-943-9574
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index