What is a typical approach to conducting a survey-based psychological experiment?

There seems to be a host of issues involved; some that seem particularly important are:

  • the appropriate design of the experiment,
  • ethical approval,
  • securing funding.

There seems to be a number of books written about the topic, and it's hard to know where to begin. I'm largely after an outline of what one needs to know in order to go from nothing to having successfully conducted a survey-based experiment.


In a previous question, I received an answer that suggested conducting a survey-based experiment. Assuming (once I do a thorough literature review) I don't find any other answer to the question, I'd be interested in implementing the experiment (partly to answer the question itself, and partly for professional development). I'm a scientist, but not in a discipline where surveys would typically arise. As such, I have no experience in conducting a survey-based experiment.

  • $\begingroup$ Are you more interested in the general issue of conducting such research, or how to do specific research on comparing attractiveness ratings for different types of objects as per the high heels question? $\endgroup$ Commented May 24, 2012 at 11:59
  • 2
    $\begingroup$ I do not think that there is a general answer to your question, as they are dependent on your research topic, the institution you conduct your research in, and the methods you want to use. $\endgroup$
    – H.Muster
    Commented May 24, 2012 at 12:00
  • $\begingroup$ At this stage, I'm interested in the general question about how to conduct survey experiments (the motivation being the previous question). If it's impractical for me to conduct such an experiment (maybe due to something unforeseen), it's probably not worth my while going into detail researching the topic. $\endgroup$
    – anonymous
    Commented May 24, 2012 at 12:13
  • 2
    $\begingroup$ As I said, it depends on what you are intending. You can do the high-heels study by just creating an internet site with your pictures and some kind of voting system. But you can also do it in a highly controlled lab with carefully designed stimuli, and perfectly balanced groups of participants. The first study would be much cheaper and easier, whereas the second one would give results that are probably more reliable and easier to publish. Nevertheless, the first study might be good enough to give you a first impression. $\endgroup$
    – H.Muster
    Commented May 24, 2012 at 13:07
  • $\begingroup$ I think the securing funding part is a completely separate question. I don't see what would be special about a survey-type experiment in terms of securing funding (except that you need much less funding). The appropriate-design of the experiment cannot be answered without knowing which experiment you want to conduct and what you expect to measure. $\endgroup$ Commented May 24, 2012 at 16:50