Structuring a talk or paper

The broad goal of your talk or paper is to convey to your audience your novel contribution to an existing literature. The following format often works quite well toward that end. One reason it works well is that it subordinates the data you have to the question you're asking, and to prior work already done on that question. Consider this a convenient, ready-made, and thoroughly road-tested template.
  1. The general question. In this section, state the general overall question that you have chosen to pursue, and explain briefly why this question is important. For a talk, this can be as short as 2 or 3 sentences, but they're important ones. E.g. "My research asks whether grammatical representations are affected by linguistic usage. This question has attracted attention recently, because it opposes two very different views of the nature of grammar."
  2. Literature review. What existing theories attempt to answer the question raised in Section 1? Briefly describe the major theoretical positions in the literature, and existing evidence that supports or challenges them. Again, for a short talk, this can be brief but it must be there.
  3. Open specific question. In what sense is the debate over the general question from Section 1 not yet resolved - i.e. what important specific question is left open by the literature you have just reviewed in Section 2?
  4. Research. Present your research, or research plan, for answering the open specific question from Section 3. Please be concrete about methods used. Before showing us any data, describe what specific pattern of possible results would answer the open question from Section 3, and how. Then you can present your actual results (if you have them), which will provide an answer to the question.
  5. Discussion and conclusions. Specify what one may conclude, given the outcomes from your research (or, if proposed research, from different possible outcomes). Make sure you explain the broader meaning of the possible results and how they support or challenge existing theories.
There are some sorts of research where this template may not apply readily - for instance, descriptive work on an endangered language, where the point is not so much to contribute to an ongoing theoretical dispute, but instead to capture a language before it vanishes forever. In such cases, the "literature review" section can be given over to other sorts of relevant background, e.g. where the language is spoken, how many speakers, why it is of particular importance. But even in such cases, do try to frame the talk in terms of broad goals acccessible to everyone: "Analyzing endangered languages is important because...", then gradually narrow in to the specifics, and finally broaden back out to end on a suitably general note.