I have quite a bit of experience with experiments – they work, if they match what you are trying to find out. For example, I recently compared what people would think of others if they were told someone liked ‘a wide range of music’ or ‘a narrow range of music’. Participants in either group did not know there was another group. I found that when people thought someone else enjoyed lots of musical styles they were rated positively, more open-minded, etc. It showed how much of a judgment we can make about someone based on something as simple as your favourite music. This is important as discussing favourite music is one of the first things people do when they first meet. So use this one to your advantage, but keep it between us! Tell people you listen to lots of music – they will think more of you! Oh how judgmental we can be.
When conducting experiments I try and keep things simple – just two groups if possible – and ensure no-one in either group knows there’s another group. I also try and keep what I am measuring relatively simple – this all helps with the maths involved when computing differences using statistics.
With any study it’s important to be quite focused. Try and do just one thing, basically. Then if your findings are ‘significant’ – ie a change did occur – you can say with some confidence that it was as a result of the experiment. Getting too ambitious and including lots of different elements can make it harder to actually explain the findings, no matter what they are.