A suggested bottom-up workflow for calibrating your research questions to the data you actually collected. This is useful if you find that you don’t have clear research questions when interpreting the results.
Research questions are shapeshifters
Throughout your paper, you answer questions. This starts in the Results, where you answer questions about the data. In the Discussion, you answer questions about the results. And in the Implications, you answer questions about the conclusions.
These question shape-shift as you move through sections, but must remain logically connected or the paper falls apart.
Think of it as a question family, not a single question:
- Measurement question (about the data; answered in Results section): “How does 20 µg/L nitrate affect growth in juvenile Atlantic salmon?”
- Findings question (about the findings/results; addressed in Discussion section): “How does nitrate affect juvenile Atlantic salmon?”
- Significance question (about the conclusions; addressed in Implications section): “How does pollution affect fish development?”
The rule: You must be able to draw a straight line from data → result → conclusion without logical breaks1. If your Discussion claims “pollution threatens global fisheries” but your Results only answered “Tank 3 grew slower”, you are missing a logical step in your reasoning.
How to calibrate your questions
Step 1: List your research questions
List the questions you answered, writing them such that they contain the specific variables you measured.
For example:
❌How does caffeine consumption affect reaction speed in students? ✅How does caffeine consumption (g/day) affect reaction speed (seconds) in master students at Utrecht University?
This forces you to think clearly about what questions you are actually able to answer directly with your data.
Replace vague words with specific ones
Vague words: an effect, the effect, an influence, to affect, to influence, to impact; some, most, many, few, several, a considerable amount of, a lot of, several; it, that, those, which, the latter, these, there, ‘was altered
❌ Many studies suggest an effect of diet on health. ✅Several longitudinal studies indicate that high dietary sugar intake increases the risk of type 2 diabetes in adults.
Step 2: Rewrite your questions in more general terms
Use the super-specific question from step 1, and replace the specific variables with more general terms.
For example:
How does caffeine consumption (g/day) affect reaction speed (seconds) in master students at Utrecht University?
becomes:
How does caffeine consumption affect cognitive performance in Dutch university students?
becomes:
How does caffeine consumption affect cognitive performance in master students globally?
Make sure you can defend the link between each jump.
For example, jumping from reaction speed to academic performance could be too far-fetched.
From specific to general
What makes a question good for writing? A question should be specific enough to be answerable with an interesting argument.
- If it’s too specific, you can can write the answer in one sentence.
- If it’s too broad, your question requires an entire book to answer.
Footnotes
-
This type of reasoning is called abstraction, and is clearly depicted on the abstraction ladder, a term coined by Hayakawa, S.I. (no date) Language In Thought And Action and further explained in Seabury, M.B. (1991) ‘Critical Thinking via the Abstraction Ladder’, The English Journal, 80(2), p. 44. Available at: https://doi.org/10.2307/818752
↩