Dear Donald McLaren Thank you for your reply! I make the contrast according to your method is following,but I want to make interaction between factor 1 (A,B)and factor 2(C,D,E), gender (M,F) and one continuous variable (age) as covariates, the following contrast:
2 -2 0 -2 2 0 2 -2 0 -2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 -2 0 -2 2 0 2 -2 0 -2 2 0 0 0 0 0 0 0 0 0 0 0 0 is correct? Thanks! Bo Xiang At 2012-11-21 04:07:37,"MCLAREN, Donald" <mclaren.don...@gmail.com> wrote: >On Tue, Nov 20, 2012 at 12:56 PM, Douglas N Greve ><gr...@nmr.mgh.harvard.edu> wrote: >> Thanks Donald. Is this the standard way to do this? I had used 8 rows >> instead of 4 with the difference being that 8 rows gives you an opportunity >> to look for an effect in males OR females. > >Yes. Having 8 rows would tell you if you have an interaction between >factor 1 and 2 in either males or females. My 4 rows only tell you if >the interaction exists. Technically speaking, one would run the >three-way interaction first. If nothing existed then you do the >two-way interaction as I suggested. If there is a three-way >interaction, then you would use Doug's approach of the interaction in >either males or females. > >If there is an effect in both >> males and females but the effects go in opposite directions, then the 4 row >> implementation will resolve to 0 (no effect). Or am I misunderstanding >> something (again:)? > >Nope. You are right. If the male and female effects are different, >then they could cancel each other out. If you suspect this to be the >case, then you should be able to demonstrate a three-way interaction. > >> thanks! >> doug >> >> >> On 11/20/2012 01:50 PM, MCLAREN, Donald wrote: >>> >>> Bo, >>> >>> Doug asked me to chime in on your issue. Here are some points that you >>> (and others) will hopefully find useful. >>> >>> (1) Inferences are two-step process. First, you create and estimate >>> the design matrix. Every column in the design matrix accounts can >>> account for some of the variance in the data. Second, you have >>> contrasts that allow you to infer specific effects. Because the model >>> contains your covariates, you are always controlling for the >>> covariates and by extension any factor/covariate not in the contrast. >>> >>> (2) Forming contrasts is often the most difficult thing to do. I >>> assume that your three factors (1, 2, and gender) are all >>> between-subject factors. If one of them is a within-subject factor >>> please let me know and disregard the rest of the email. The final >>> F-contrast will have 4 rows (factor 1 levels-1)*(factor 2 levels >>> -1)=(3-1)*(3-1)=2*2=4 >>> >>> The following is an outline for creating contrasts: >>> (a) Start simple - difference between levels of 1 factor >>> (b) Define your null hypothesis: AO=AP=AQ >>> (c) Make it equal to 0: AO-AP=0 AND AP-AQ=0 >>> (d) Repeat for the other levels of the factor... >>> BO-BP=0 AND BP-BQ=0 >>> CO-CP=0 AND CP-CQ=0 >>> >>> (e) Now combine them AO-AP=BO-BP=CO-CP AND AP-AQ=BP-BQ=CP-CQ >>> >>> (f) Make them equal to 0: >>> AO-AP-BO+BP=0 >>> BO-BP-CO+CP=0 >>> AP-AQ-BP+BQ=0 >>> BP-BQ-CP+CQ=0 >>> >>> (g) Expand them to include gender, for example: >>> AO-AP-BO+BP=0 becomes FAO-FAP-FBO+FBP+MAO-MAP-MBO+MBP=0 >>> >>> Since the contrast now has 2 columns per level, you should divide all >>> values by 2. This will produce the correct amplitude and statistics. >>> If you leave the values as 1 and -1, then you will have an incorrect >>> amplitude, but the statistics will still be correct. >>> >>> (h) Fill in the respective columns of your design matrix. >>> >>> (3) The degrees of freedom are defined based on the rows of the >>> F-matrix and the number of rows in the design matrix. The F-test has a >>> numerator and denominator degrees of freedom. F(n,d). >>> >>> Best Regards, Donald McLaren >>> ================= >>> D.G. McLaren, Ph.D. >>> Research Fellow, Department of Neurology, Massachusetts General Hospital >>> and >>> Harvard Medical School >>> Postdoctoral Research Fellow, GRECC, Bedford VA >>> Website: http://www.martinos.org/~mclaren >>> Office: (773) 406-2464 >>> ===================== >>> This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED >>> HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is >>> intended only for the use of the individual or entity named above. If the >>> reader of the e-mail is not the intended recipient or the employee or >>> agent >>> responsible for delivering it to the intended recipient, you are hereby >>> notified that you are in possession of confidential and privileged >>> information. Any unauthorized use, disclosure, copying or the taking of >>> any >>> action in reliance on the contents of this information is strictly >>> prohibited and may be unlawful. If you have received this e-mail >>> unintentionally, please immediately notify the sender via telephone at >>> (773) >>> 406-2464 or email. >>> >>> >>> On Tue, Nov 20, 2012 at 8:17 AM, xiangbo_2010<xiangbo_2...@126.com> >>> wrote: >>>> >>>> Dear doug >>>> Thank you for your reply! >>>> +AOM -BOM -APM -BPM 0 0 0 0 0 >>>> +AOM -BOM 0 0 -AQM +BQM 0 0 0 >>>> +AOM 0 -APM 0 0 0 -COM +CPM 0 >>>> +AOM 0 0 0 -AQM -COM 0 +CQM >>>> there should be use 1 -1 or 0.5 -0.5? whether the -BPM should be change >>>> BPM? >>>> Thanks >>>> >>>> Bo Xiang >>>> >>>> >>>> >>>> >>>> >>>> >>>> At 2012-11-19 07:23:31,"Douglas Greve"<gr...@nmr.mgh.harvard.edu> wrote: >>>> >>>> >>>> Hi Bo, you can think of the Ftest as a logical 'OR' between the t-test >>>> contrasts indicated in each row. Each row is a difference of differences, >>>> so >>>> >>>> 1. (A-B)om - (A-B)pm --> Does the difference between A and B differ >>>> between >>>> groups O and P for Males? >>>> 2. (A-B)om - (A-B)qm >>>> 3. (A-C)om - (A-C)pm >>>> 4. (A-C)om - (A-C)qm >>>> 5. (A-B)of - (A-B)pf --> Does the difference between A and B differ >>>> between >>>> groups O and P for Females? >>>> 6. (A-B)of - (A-B)qf >>>> 7. (A-C)of - (A-C)pf >>>> 8. (A-C)of - (A-C)pf >>>> >>>> I've put together the first 9 columns of the first 4 rows. The last 9 >>>> columns are all 0s. For the last for rows, the 0s and below matrix are >>>> swapped to give you the same for the females >>>> >>>> doug >>>> >>>> AOM BOM APM BPM AQM BQM COM CPM CQM >>>> >>>> -------------------------------------------------------------------------- >>>> +AOM -BOM -APM -BPM 0 0 0 0 0 >>>> +AOM -BOM 0 0 -AQM +BQM 0 0 0 >>>> +AOM 0 -APM 0 0 0 -COM +CPM 0&! nbsp;&nb sp; >>>> >>>> +AOM 0 0 0 -AQM -COM 0 +CQM >>>> >>>> >>>> >>>> On 11/17/12 9:21 PM, xiangbo_2010 wrote: >>>> >>>> Hi Freesurfer experts, >>>> >>>> I'm very sorry to bother you, but I am very confused with the following >>>> questions: >>>> >>>> My experimental design includes three discrete factors: factor 1 with >>>> three >>>> levels (A,B,C ); factor 2 with three levels (O,P,Q); gender (F, M), and >>>> one >>>> covariate. >>>> >>>> So I can get 18 classes: FAO, FAP,FAQ,FBO,FBP,FBQ,FCO,FCP,FCQ,MAO, >>>> MAP,MAQ,MBO,MBP,MBQ,MCO,MCP,MCQ. I want to perform the interaction >>>> between >>>> factor 1 and factor 2 regressing out the effect of gender and one >>>> covariate, >>>> but I don't know the rules for setting the contrasts for the F-test. The >>>> contrast matrix I used is: >>>> >>>> 1 1 1 -1 -1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >>>> 0 >>>> >>>> 1 -1 0 1 -1 0 1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >>>> 0 >>>> >>>> 1 1 1 0 0 0 -1 -1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >>>> 0 >>>> >>>> 1 0 -1 1 0 -1 1 0 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >>>> 0 >>>> >>>> 0 0 0 0 0 0 0 0 0 1 1 1 -1 -1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >>>> 0 >>>> >>>> 0 0 0 0 0 0 0 0 0 1 -1 0 1 -1 0 1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >>>> 0 >>>> >>>> 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 -1 -1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >>>> 0 >>>> >>>> 0 0 0 0 0 0 0 0 0 1 0 -1 1 0 -1 1 0 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >>>> 0 >>>> >>>> is it correct? >>>> >>>> >>>> >>>> Any help will be very appreciated. >>>> >>>> >>>> >>>> Best wishes, >>>> >>>> >>>> >>>> Bo Xiang >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> _______________________________________________ >>>> Freesurfer mailing list >>>> Freesurfer@nmr.mgh.harvard.edu >>>> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer >>>> >>>> >>>> >>>> >>>> >>>> _______________________________________________ >>>> Freesurfer mailing list >>>> Freesurfer@nmr.mgh.harvard.edu >>>> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer >>>> >>>> >>>> The information in this e-mail is intended only for the person to whom it >>>> is >>>> addressed. If you believe this e-mail was sent to you in error and the >>>> e-mail >>>> contains patient information, please contact the Partners Compliance >>>> HelpLine at >>>> http://www.partners.org/complianceline . If the e-mail was sent to you in >>>> error >>>> but does not contain patient information, please contact the sender and >>>> properly >>>> dispose of the e-mail. >>>> >>> >> >> -- >> Douglas N. Greve, Ph.D. >> MGH-NMR Center >> gr...@nmr.mgh.harvard.edu >> Phone Number: 617-724-2358 >> Fax: 617-726-7422 >> >> Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting >> FileDrop: www.nmr.mgh.harvard.edu/facility/filedrop/index.html >> Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ >> >_______________________________________________ >Freesurfer mailing list >Freesurfer@nmr.mgh.harvard.edu >https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.