Hi
I'm currently writing my masters thesis and i'm analyzing data from a school-based intervention study. The statistical analysis is conducted with STATA 17.
We want to examine whether the implementation quality of a school-based intervention has an effect on the effectiveness of the intervention.
As i just started to work with mixed models i'm not sure whether my model is correct.
My statistical approach is adopted from a study by Humphrey et al. (2018): https://doi.org/10.1007/s11121-017-0802-4
I further rely on the book by Twisk (2019). You can access the relevant part here: https://1drv.ms/u/s!AggQKHp7NGMFg6NMkBiK-rjcZ-EodQ?e=23xJmS
Stress measurement:
z_k_sskj_vul1: vulnerability to stress at time 1 (continuous, z-transformed)
z_k_sskj_vul2: vulnerability to stress at time 2 (continuous, z-transformed)
Implementation Quality:
resp_high: high responsiveness of children (1 if high, otherwise 0)
resp_med: medium responsiveness of children (1 if medium, otherwise 0)
resp_low: low responsiveness of children (1 if low, otherwise 0)
qual_high: high responsiveness of children (1 if high, otherwise 0)
qual_med: medium responsiveness of children (1 if medium, otherwise 0)
qual_low: low responsiveness of children (1 if low, otherwise 0)
--> The implementation quality is measured on the classroom level, stress is measured on the students level.
I choose a two-level approach with students on level 1 and classrooms on level 2.
First i checked the ICC in order to assess the between-group variable. Here i'm not sure whether to include z_k_sskj_vul1 in the model.
## unconditional mdel z_k_sskj_vul2 (check for between-classroom variance)
mixed z_k_sskj_vul2 || classroom:
estat icc
--> output: https://1drv.ms/u/s!AggQKHp7NGMFg6NMkBiK-rjcZ-EodQ?e=23xJmS
Second i built this model in order to assess whether the effect differs for high or medium implementation quality in comparison with low implementation quality. I included z_k_sskj_vul1 to control for baseline measurements.
## mixed model to examine implementation quality
mixed z_k_sskj_vul2 resp_high resp_med qual_high qual_med z_k_sskj_vul1 || classroom:
--> output: https://1drv.ms/u/s!AggQKHp7NGMFg6NMkBiK-rjcZ-EodQ?e=23xJmS
I would be really grateful if someone could just check if my model is correct.
Thank you
David
I'm currently writing my masters thesis and i'm analyzing data from a school-based intervention study. The statistical analysis is conducted with STATA 17.
We want to examine whether the implementation quality of a school-based intervention has an effect on the effectiveness of the intervention.
As i just started to work with mixed models i'm not sure whether my model is correct.
My statistical approach is adopted from a study by Humphrey et al. (2018): https://doi.org/10.1007/s11121-017-0802-4
I further rely on the book by Twisk (2019). You can access the relevant part here: https://1drv.ms/u/s!AggQKHp7NGMFg6NMkBiK-rjcZ-EodQ?e=23xJmS
Stress measurement:
z_k_sskj_vul1: vulnerability to stress at time 1 (continuous, z-transformed)
z_k_sskj_vul2: vulnerability to stress at time 2 (continuous, z-transformed)
Implementation Quality:
resp_high: high responsiveness of children (1 if high, otherwise 0)
resp_med: medium responsiveness of children (1 if medium, otherwise 0)
resp_low: low responsiveness of children (1 if low, otherwise 0)
qual_high: high responsiveness of children (1 if high, otherwise 0)
qual_med: medium responsiveness of children (1 if medium, otherwise 0)
qual_low: low responsiveness of children (1 if low, otherwise 0)
--> The implementation quality is measured on the classroom level, stress is measured on the students level.
I choose a two-level approach with students on level 1 and classrooms on level 2.
First i checked the ICC in order to assess the between-group variable. Here i'm not sure whether to include z_k_sskj_vul1 in the model.
## unconditional mdel z_k_sskj_vul2 (check for between-classroom variance)
mixed z_k_sskj_vul2 || classroom:
estat icc
--> output: https://1drv.ms/u/s!AggQKHp7NGMFg6NMkBiK-rjcZ-EodQ?e=23xJmS
Second i built this model in order to assess whether the effect differs for high or medium implementation quality in comparison with low implementation quality. I included z_k_sskj_vul1 to control for baseline measurements.
## mixed model to examine implementation quality
mixed z_k_sskj_vul2 resp_high resp_med qual_high qual_med z_k_sskj_vul1 || classroom:
--> output: https://1drv.ms/u/s!AggQKHp7NGMFg6NMkBiK-rjcZ-EodQ?e=23xJmS
I would be really grateful if someone could just check if my model is correct.
Thank you
David
