https://communities.sas.com/t5/Statistical-Procedures/unequally-spaced-time-interval/m-p/55605#M2547 Can unequally space time interval be specified by models such as AR, ARMA or TOEP?<BR /> or I need to use SP(Pow)? If so, could anyone show me how to do it.<BR /> <BR /> Another more fundamental question , is it necessary to include time effects anytime conducting longitudinal or time series data if time is not the interest.<BR /> <BR /> Thank you Sat, 23 Apr 2011 14:59:17 GMT cvl 2011-04-23T14:59:17Z https://communities.sas.com/t5/Statistical-Procedures/unequally-spaced-time-interval/m-p/55605#M2547 Can unequally space time interval be specified by models such as AR, ARMA or TOEP?<BR /> or I need to use SP(Pow)? If so, could anyone show me how to do it.<BR /> <BR /> Another more fundamental question , is it necessary to include time effects anytime conducting longitudinal or time series data if time is not the interest.<BR /> <BR /> Thank you Sat, 23 Apr 2011 14:59:17 GMT https://communities.sas.com/t5/Statistical-Procedures/unequally-spaced-time-interval/m-p/55605#M2547 cvl 2011-04-23T14:59:17Z https://communities.sas.com/t5/Statistical-Procedures/unequally-spaced-time-interval/m-p/55606#M2548 If you collect data over time on the same experimental units, you need to account for the repeated measure in the statistical model, whether you care about time or not. If times are not equally spaced, you should normally use one of the of the "SP()()" structures in the repeated statement. You need a variable for time (let's call it time). Then, you could use <BR /> repeated / sub=subject type=sp(pow)(time);<BR /> where subject defines the experimental unit. <BR /> <BR /> If the times are close to being equally spaced, it is probably a reasonable approximation to use ar(1), or one of the other ones. Mon, 25 Apr 2011 19:09:36 GMT https://communities.sas.com/t5/Statistical-Procedures/unequally-spaced-time-interval/m-p/55606#M2548 lvm 2011-04-25T19:09:36Z