A good abstract should clearly state what you will research, how you will do it, and what policy relevance your project will have (this should include for whom it will be relevant). You should be as specific as possible (e.g., if you plan to cover 3-5 blocks, state this).
You should clearly state the methods you are planning to deploy. Refer to the usual textbooks (e.g., for quantitative, Counterfactuals and Causal Inference Methods and Principles for Social Research or Mostly Harmless Econometrics: An Empiricist's Companion; for qualitative, The Oxford Handbook of Qualitative Research).
Cite literature in your abstract. This will signal to us whether you are familiar with the relevant literature in the field. You should focus on high-quality literature from reputed peer-review academic journals or international academic publishers (such as Cambridge University Press or Oxford University Press).
For senior researchers and institutions only: you should also elaborate on your links to government officials and how you plan to disseminate your findings to relevant stakeholders.