State the question your paper will address. This may be stated as a formal hypothesis, but should be concise rather than detailed. The first sentence of an abstract is often also the topic sentence for the first paragraph of the paper. Say what you were researching and why you went looking for it.
Summarize in a sentence or two your review of other papers and books you used as background, with something like "Longitudinal studies exploring this question across the last quarter of the 20th century confirmed the hypothesis, but the latest of those studies reported a third strongly correlated variable that has not yet been thoroughly explored." Save the details of who did those studies, how many subjects they studied, and their statistics for the body of your paper unless your literature review revealed a significant disagreement among other researchers.
Review the methodology for your research, but again save the details. If you collected original data, state simply how many subjects were in the study and whether you collected qualitative data, designed your own survey or assessment instrument, or employed a previously validated instrument. If you analyzed an existing database, identify it as it is known in the discipline. For example, if your method was a systematic comparison of those 15 longitudinal studies conducted over 25 years, state in your abstract how you selected the studies to be included and how you resolved inconsistencies among them.
Focus a brief statement of your findings on what your study adds to what was known before about this question, even if it's only to say that you were able to replicate earlier studies or to arrive at similar findings using a different methodology or a new data set. If you have one or two very strong findings you can state them, but put the detailed statistical analysis in the paper itself.
Restate your hypothesis as either confirmed or nullified. The abstract might say no more than that, since the conclusion of the paper will briefly summarize the findings. You might, however, include in the abstract a brief "teaser" for whatever next research step may be motivated by your findings.
Write the paper, using your first-draft abstract as an outline. Make notes as you go along if you find better ways to organize some of the sections; they will help you revise the abstract to match the paper.
Read the paper over with the draft abstract in your hand or on the screen of your computer. Note any inconsistencies between the two by marking changes you will need to make to the abstract so it reflects the paper accurately. Smooth out the abstract and read the two side by side at least once more.