I have 4 timepoints of data - Students were assessed in fall and spring over two school years - with programming during the school year.
Examination of descriptive data shows no change in any variable from time 2 (spring year 1) to time 3 (fall year 2) – as there is no programming at this time, we would also not expect any change at this time. I’m wondering if using Time as just a continuous variable (1-4) will not give us as much information about change than if we tried to examine change in year 1 and change in year 2 separately. My original coding is below, but I’m not sure if this is correct or would be useful:
Time TimeBLC Time34C
1 0 0
2 1 0
3 1 0
4 1 1
Hi,
I'm assuming that TimeBLC and Time34C are two alternative methods for coding time. If so, then TimeBLC tests the hypothesis that the mean at Time 1 is no different than the combined mean of Times 2-4. Time34C tests the hypothesis that the combined mean of Times 1-3 are no different than the mean at Time 4.
Since you think the means should show discontinuous change, one option would be to treat time as a factor and create dummy variables:
Time Dummy1 Dummy2 Dummy3
1 0 0 0
2 1 0 0
3 0 1 0
4 0 0 1
Then using the set of variables, Dummy1 - Dummy3 would represent time in a categorical fashion. This particular coding uses Time 1 as the reference point, so the intercept would represent the mean at Time 1. The coefficients for Dummy1-Dummy3 would then represent mean differences between later times and Time1. You can set the reference time to whichever time you find more interesting.
It seems you have to create your own dummy variables for HLM 6:
http://www.ats.ucla.edu/stat/hlm/seminars/hlm6/outline_hlm_seminar.pdf
(see slide 18)
but that could easily be done outside of HLM, say in SPSS.
I would suggest the following:
in SPSS, create 4 variables {name them T1, T2, T3, T4}
in the raws, each student score in the programming test will be coded 4 times {assuming you had the 4 scores for each student}
you can run Repeated measure ANOVA test.
Repeated measures ANOVA is the equivalent of the one-way ANOVA, but for related, not independent groups, and is the extension of thedependent t-test. A repeated measures ANOVA is also referred to as a within-subjects ANOVA or ANOVA for correlated samples.
The following link has some detials: Good luck
https://statistics.laerd.com/statistical-guides/repeated-measures-anova-statistical-guide.php
Hi,
I'm assuming that TimeBLC and Time34C are two alternative methods for coding time. If so, then TimeBLC tests the hypothesis that the mean at Time 1 is no different than the combined mean of Times 2-4. Time34C tests the hypothesis that the combined mean of Times 1-3 are no different than the mean at Time 4.
Since you think the means should show discontinuous change, one option would be to treat time as a factor and create dummy variables:
Time Dummy1 Dummy2 Dummy3
1 0 0 0
2 1 0 0
3 0 1 0
4 0 0 1
Then using the set of variables, Dummy1 - Dummy3 would represent time in a categorical fashion. This particular coding uses Time 1 as the reference point, so the intercept would represent the mean at Time 1. The coefficients for Dummy1-Dummy3 would then represent mean differences between later times and Time1. You can set the reference time to whichever time you find more interesting.
It seems you have to create your own dummy variables for HLM 6:
http://www.ats.ucla.edu/stat/hlm/seminars/hlm6/outline_hlm_seminar.pdf
(see slide 18)
but that could easily be done outside of HLM, say in SPSS.
- Badges
- Science topic
- Similar topics
- Psychology
- Psychometrics