Having in mind the sorts of things that drive research can help to identify new problems to work on and can help to frame the rationale for a project. Of course, all original research aims to close a “gap” in the literature, extend the literature or clarify something in the literature. But close what gaps, extend where, clarify what?

Generally speaking, original research aims to do one or more of the following things:

Describe something

Description would probably also be coupled with explanation (e.g. "How are gorilla social groups in the wild organised?" or "How well does post-revolutionary Cuban cinema reflect the revolution’s ethos of gender equality?"). So an argument would have to be presented as to why existing descriptions are deficient, or to claim that they are non-existent but nevertheless of interest.

Explain something

For example, "Why is the universe currently expanding at an ever increasing rate when one would expect gravity to slow the rate of expansion?" or "Why do students drop out of university?". In cases such as these, an argument would need to be put forward as to why current explanations are deficient and of what benefit there would be in having further explanation.

Test an important aspect of a current theory

For example, "General relativity predicts that rotating gravitational objects drag spacetime around with them, but is this true?" Testing this idea was one of the goals of the Gravity Probe B project, so an argument as to why this needs testing, or testing more stringently, needs to be made.

Replicate an important study

This is common in medical research where confounding factors can bias results, which leads to many results being treated with suspicion until proven by multiple studies. For such research, reasons for having less than full confidence in existing results (perhaps because of limitations in the size or representativeness of the sample) would need to be presented. If the issue is that you think the existing results might be out of date (e.g. if more than 20 years old, social research in a developing country might be out of date if that country has seen significant development and/or exposure to different cultures in the last 20 years), this might also provide a reason for redoing a study.

Retest a hypothesis

Using new or improved methodological procedures.

Resolve a conflict in the literature

For example, "Are certain types of student errors in physics the result of their having (incorrect) preconceptions about physical principles, a result of ad hoc activation of intuitive rules, or a combination of both depending on the student and the question?". When the goal is to resolve a conflict, the nature of and reasons for the conflict would need to be explained.

Solve a problem

For example (from the early days of flight), "How do airframes need to be modified so that aircraft can fly faster than the speed of sound without disintegrating?". For this case, the nature and importance of the problem would need to be explained, together with why existing solutions are not completely satisfactory.

Improve something

For example, working out how to make more efficient photovoltaic cells or faster processors for computers. Why those improvements would be worth striving for would need to be explained.

Evaluate something

For example, "The government believes that their policy X should help with the problem of Y, but how well has this been realised in practice?".