Public Health Research Questions

It names an exposure, an outcome and a place, and it claims only what the design can carry. "How is the spatial variability of acute pediatric asthma related to PM2.5 concentrations and the amount of vegetated cover?" pairs two exposures with one outcome, and the word spatial tells a reader the analysis is geographic before the methods chapter does. Related to, rather than causes, is the honest promise for observational data. Copy that shape: two named things, one direction of interest, and a population or a place you can actually get records for.

Decide before you write the sentence, because it changes the data you need. A question comparing women living in areas of high disadvantage with women in areas of low disadvantage is comparing areas, not people, and the obvious examiner question becomes how disadvantage was defined. The level of measurement works the same way. Access to a resource can be counted per person, per household or per district, and a question that says which one removes the commonest source of confusion in a prevalence chapter.

In the question, or at least before you collect anything. Where the grouping is a social category such as race, income, migration status or rurality, the question has already committed you to a classification, and an examiner will usually ask which one and why. Use the definition your data source already uses where you can, since a registry or a national survey has a coding scheme you will have to live with anyway. Inventing your own categories in the results chapter is what turns a finding into an argument about your categories.

Then the question changes, not the data. Administrative and registry files exist for a defined place over a defined period, and they hold the variables somebody else decided to collect, so read the codebook before you write the sentence. Students lose months writing a question about something the extract never recorded. If the variable you need isn't there, either find a source that has it or ask the nearest question the file can answer honestly.

Yes, as a pair of questions rather than one blurred sentence. A study can measure the change in active travel and outdoor play between two phases, then ask community leaders and children what lay behind it. Each half is answerable on its own, which is what makes the pair work: the numbers say what happened, the interviews say what the people there think it was about, and neither is asked to do the other's job.

Usually a determination rather than a full review, and the board makes that call, not you. In the US the exemption for secondary research turns on conditions you can check: the information is already publicly available, or it is recorded so identities cannot readily be ascertained, you do not contact anyone, and you will not re-identify them. Ask for the letter anyway, because a journal may want it later. A registry extract rather than a public download usually means a data use agreement too, and other countries route research ethics differently, so check your own committee's route.

Only if your design has a comparison that isolates the policy. A question asking whether foregoing care among Georgia residents is related to the passage of the Affordable Care Act, with year standing in as the proxy, has that relational wording doing real work. One population measured before and after leaves you unable to rule out everything else that happened in those years. Write associated with, and put the causal argument in your discussion where it can be qualified.

Yes, and in applied public health programs it is often the expected shape. It becomes a thesis rather than a report when the question can be answered against a stated standard: whether the meals served to learners constitute a balanced meal, whether hygiene practice met a named regulation. The version that gets sent back is the vague one, asking whether the program is working. Programs differ on what they will accept, so read your own program guidance before you commit to an evaluation.

Almost always, twice. Your own institution reviews it, and so does a body in the country where you collect, which in many places is a national health research council with its own fee and forms. That second clearance runs on the host country's timetable rather than your university's, and it is often the longest step in the whole project. What it takes differs from country to country, so ask both committees what they want while the proposal is still in draft.

Recent enough that your conclusion still holds, which is a judgment about the topic rather than a fixed rule. National surveys and registries run two or three years behind by the time they are released and nobody objects to that. What does draw an objection is a pre-pandemic file used to describe current service use, or pre-policy data used to comment on a policy. Name the years inside the question and the gap becomes a stated delimitation rather than something a reader catches you hiding.