repeated measurement multilevel modeling: research design

[nasirihakimeh]

Hi, this is Hakimeh.
I am a new learner of repeated measurement multilevel modeling.
Kindly, I seek your advice to design the following research appropriately.

Research objectives:
1) To explore the variation of patterns of kids’ smartphone usage during typical everyday life in comparison to travel time
2) To explore the association of the above-mentioned pattern to smartphone addiction or non-addiction of parents of those kids

Research design
Stage 1: Using a scale I will categorize parents into two groups of Smartphone Non-addicted Parents vs Smartphone Addicted Parents
Stage 2: another scale will measure patterns of kids’ smartphone usage in two occasions:
Occasion 1: typical everyday life (I will repeat asking the same questions during one week-days)
Occasion 2: travel time (I will repeat collecting asking the same questions during a family travel period which can vary from 1 day to 5 days. Please note that 5 days is the average travel time of the families in my target society)

I need your advice on:
1. Regarding the sample size, how many parents should be sampled?
2. I have visualized the structure of the prospective data that will be collected for the above-mentioned research idea in Figure 1 vs Figure 2. Which one is correct?

Figure 1

figure 1.PNG (35.24 KiB) Viewed 15446 times

Figure 2

figure 2.PNG (43.83 KiB) Viewed 15446 times

Warm regards,
Hakimeh

[billb]

Hi Hakimeh,
Not sure I fully follow your research but I'll try and answer - your diagrams are a bit weird as I would have just put parent1, parent 2 etc. and then you would use whether they are addicted or not as a covariate - otherwise it looks like you only have 2 parents. I was unclear of what the response variable measured on the occasion was? and definitely I would have repeats within kids within parents like the left hand of figure 1.
I wasn't sure why the specific 2 occasions but guess you could again put a covariate for occasion type. Sample size will depend on correlation between measures, desired effect size to detect etc. See http://www.bristol.ac.uk/cmm/software/mlpowsim/ and the manual and talk slides there for some ideas.
Best wishes,
Bill.

[nasirihakimeh]

Dear Bill,

Your response is well received.
I appreciate it. Many thanks.

Warm Regards,
Hakimeh