Re: Mixed syntax for repeated measure at level II

Posted by Ryan on Sep 03, 2016; 5:54pm
URL: http://spssx-discussion.165.s1.nabble.com/Mixed-syntax-for-repeated-measure-at-level-II-tp5733009p5733047.html

Without a student indicator, it wouldn't be possible to capture student variance.

Someone mentioned growth curve models, but one would really need more than 2 time points to measure growth in a meaningful way. That said, the idea of a growth curve model with these type of data piqued my interest. Below would be a reasonable starting point to fit a growth curve model for these type of data assuming at least three time points:

Y = mu + time_within_school_ij + residual_ijk 

For a given school, the time effects have the following covariance structure
        | time_in_school_1j |                                     |  1   rho  rho^2 |
    Var | time_in_school_2j |  =  sigma_time^2                    |      1      rho |
        | time_in_school_3j |                                     |               1 |

residual variance = variance among the replicate observations in the same time at the same school

variance  for subject "school" is "sigma_time^2" = variance among the effects of time within a given school

first-order autoregressive correlation for subject "school" is "rho" = correlation between time within school effects that are adjacent in time

Ryan

On Sun, Aug 28, 2016 at 5:03 PM, Bruce Weaver <[hidden email]> wrote:

William, you said you cannot link students from time 1 to time 2.  I gather
this means that there are students (possibly a substantial number) who have
data for both time points, but you lack the unique student IDs required to
link the two data points.  Is that correct?  If so, I imagine there are also
some students who have data for only one year.

Thanks for clarifying.




William Dudley wrote
> I need to look at change over time  (year 1 year 2) in Math scores
> measured at the student level with students nested within Schools.
>
> I have the same 13 schools at both time points
> However I cannot link students from time one to time two.
> The number of students within a school varies both within a year and
> across
> schools
> and within a school across years.
>
>
> Is there a way to model the effect of time
>  in SPSS Mixed so that I can capitalize on the fact that schools are
> repeatedly measured.?
>
> Thanks in advance
>
>
>
>
>
> --
> William N Dudley, PhD
> President
> Piedmont Research Strategies, Inc
>
> =====================
> To manage your subscription to SPSSX-L, send a message to

> LISTSERV@.UGA

>  (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD





-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Mixed-syntax-for-repeated-measure-at-level-II-tp5733009p5733010.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD