Can I combine data from different sources?

#1
Hi all, I'm very novice at statistics but part of my course is to complete a small research project. My supervisor has said that she can't help me with statistics so I'm at a loss and would be very grateful for any help anyone can offer here.

My research question is whether bullying experiences in school impact on body dissatisfaction and engagement in body change behaviors in later life.

I have posted a survey in two locations: one to undergrad psychology students and the other to a convenience sample on social networking websites. I have about 50 participants from each sample. For the sake of power, I would ideally like to combine the two samples.

The samples do seem to differ on demographic measures. The undergrad students are younger, have completed less education, are predominantly single, and are more likely to work casual jobs compared to the participants from the social networking website.

My question is: what do I need to look at to determine if these differences matter for my research question?
 
#2
Hey yokolatte.

You can indeed combine data from multiple sources and provided the attributes are representative in both areas (in terms of what they measure and what it means), then the area that deals with aggregation of data is known as meta-analysis.

When you combine data from different sources you need to take into account covariance and intra-class correlation (ICC). If you consider data to be truly independent statistically, then these will be zero and you will simply be left with a weighted sum of statistics.

If you want to understand meta-analyses then take a look in statistics books or journals on the topic (especially those in bio-statistics which often have clinical trials across many centres).

If you have specific questions, I can do my best to answer them.