Read previous literature - see what has been done. Where is the evidence? Where is there a lack of evidence? Hypotheses should be based in the literature - based on theory - based on experience ... You are going to need to justify how you came up with that hypothesis - so you need to base it on something - and previous research evidence is a good starting point :)
Faheemullah - Developing a hypothesis basically varies with the kind of research project conducted: qualitative and quantitative. There are several criteria that can be used to develop a hypothesis 1) A hypothesis is stated in a declarative form and not as a question 2) A hypothesis clearly states a relationship between variables 3) Does the hypothesis include both an independent and dependent variable? Are they easy to identify 4) Make sure that the hypothesis clearly defines the topic and the focus of the study. 5) Hypotheses are not moral or ethical questions but should be tested - verifiable or falsifiable 6) A hypotheses should be in the form of single tentative guess/s, good hunches 7) There are basically two types of hypothesis, namely, null hypothesis (Denoted by "HO") and alternative hypothesis (Denoted by "H1"). Personal experience, imagination & thinking, observation, scientific theory, previous study and culture are some of the common sources of hypothesis.