Firstly, look at the question, perhaps highlight the key words. What type of hypothesis are they looking for?
Are they looking for a 'null' hyopthesis? - stating there will be no difference.
Or a 'directional' or 'non-directional' hypothesis? - the direction may be suggested in the text before the question.
Perhaps the study is correlational and you must determine whether the correlation is likely to be positive or negative.
Always operationalise your variables - ie. make them measurable, this is where you describe the independent and dependent variables in terms of this study. eg. Don't just say memory - say 'number of words recalled'.
Include the word significant: significant difference, significantly higher, a significant positive correlation. This means that you believe the data will show a trend when a statistical test is used to measure it - usually with 95% confidence.
Begin with 'There will be...' and put these things together - you should have the perfect hypothesis!