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. > _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
