Statistics

This is hard to answer without context. How are they very different?

My first thought is that you shouldn’t attempt to compare data that has been collected in different ways. Doing it in different ways can lead to the data implying relationships that don’t exist - have you spoken to your lecturer about it?

So it sounds like you're doing something similar to a meta-analysis from a range of studies. First thing you need to do is carefully consider which data sets you are going to use - this will probably only be a select few of the studies you have identified. It's important that you choose studies that have similar objectives and methodologies to minimise variance within the data sets.

On a side note, I'd expect you'd be able to write quite a lot about how you select your data and why you have excluded certain data sets. You could even do a funnel plot of the studies to infer publication bias within your analysis.

When you've narrowed down your data sets, perhaps a good starting point of analysis would be to calculate the dependent variable (the complement resistance, if it needs calculating). I'd suggest something simple like calculating the average outcome and doing an ANOVA with post hoc test to see of there's a significant difference in the outcome between the studies.

Something like that, it's a systematic review. The thing is I don't have a lot of data because there isn't much out there for me to find. I have 9 different results and they're quite different. I can combine 2 sets of data together and another set of 2. The rest of the data is too different. For calculating the average outcome, if I do that for each of the results and then compare that would that be alright? And is it possible to work out if each of the data is statistically significant separately and then compare that? do they have to be done using the same test or can I use different methods? Also, I didn't know about a funnel plot, I'll definitely look into that! Thank you so much for your help! You have helped me more than anyone else! thank you!

The problem if your studies have vastly different methodologies is it will usually increase underlying inter group variance and distort / undermine credibility in your results. You'll likely know best yourself whether this will be a tolerable level of variance or if it's just so different that it severely compromises the results.

If you can't get any more data then unfortunately you'll have to proceed with what you've got. I'd suggest first an ANOVA comparing the mean complement resistance (which I'm assuming is a covariate - not a factor response?) followed by a post hoc test - most suitable type will probably be Tukey's HSD. The post hoc will reveal data groups that are statistically different, you'll be able to use your judgement to say whether this could be due to different methodologies.

I know the feeling about having to muddle through things like this on your own. I did two dissertations on applying statistical methods in proteomics - hardest thing I did at uni by far! But it was worth it, one of mine was published recently - yours could be too! Projects like this really do give you some of the best transferable skills outside of uni I'd say, so stick with it and you'll be fine