Guidelines for Asking Research Questions
I - Scientific Inquiry
Inquiry refers to the ways in which scientists study the natural world and in which new knowledge is produced.
Types of questions you might ask about the natural world:
- Questions that require value judgments, a “yes” or “no” answer.
- Questions that require describing something using already known facts to answer them.
- Questions that require making observations, then generalizations or conclusions.
A research question asks: "What relationship exists between two or more variables?" It expresses a possible, clearly stated relationship between these variables, and the question implies possibilities for an experiment to test the relationship. Research questions are often stated as hypothesis: the prediction of a relationship between two or more variables. Classical scientific methods include hypothesis forming and testing to resolve scientific questions and build theories. A hypothesis might be tested repeatedly and thus evolve into a theory. Generating and testing hypotheses provides important practice in risk taking and divergent thinking.
II - Generating your research questions
Research questions should be grounded in observations of the phenomena in habitats nearby, and reading published reports and discussion about the phenomena. Discussion can be useful in helping to develop and refine questions.
Once you have your research topic and overarching purpose, such as an investigation of nearby forested park habitat, clearly in mind, you can start framing your research questions and generate a simple model of the important factors you think directly affect the subject. Try developing several questions related to your topic, and then choose one that best fits. Rephrase your question until it is:
- analytical, requiring analysis of data
- focused and based upon previous information on the topic to prevent a superficial discussion
- somewhat explanative or predictive of expected results
In the process of carrying out your research, additional questions may be raised that in turn lead to further hypotheses.
III - Help narrowing down your question:
Research questions are often flawed in one of the following ways:
- The question is not answerable by science
- The question is not currently answerable due to technology limitations
- The question is either too expensive or unfeasible due to logistics
- The question is too broad to handle
By spending the time to create a good research question, the rest of the process will be much more manageable
- Does your question pass these tests:
- Can it be answered scientifically; by observation, experimentation, replication?
- Does the technology exist to do any of the necessary measurements?
- Do you have access to the equipment, location, have the time and other resources to adequately answer it?
- If all answers to the above questions in #1 are yes, you are probably ready to go. If all except time and resources are yes, then proceed to step 3. If your question fails any other item in the above, you will probably need to alter your question at this time.
- Refine your question using this checklist:
- Is there a clear method for finding an answer?
- Is the question genuine (not already known and not answerable with a “yes’ or “no”)?
- Is the question useful to know about?
- Is it possible to predict an outcome?
- Does the question contribute to a bigger question or other questions?
- If you answer “no” to any of the above questions in #3, you probably should modify your question. If you answered “yes” to all of the above, you are probably ready to go!
- What will you need to provide evidence for your question?
- Have you been able to find papers from peer-reviewed scientific journals that are relevant to your question?
- Are you working with or seeking professional advice from a partnering researcher or a scientist from a nearby university or other site about your question?
- Did you make your own observations at the site?
- Do you have a plan about how you will collect your own data using an established research protocol during your experiment?
- Do you have a model about your study of how your variables relate to one another that you created?
- If you answered “no” to any of the questions above, you probably should do more groundwork. If you can answer “yes” to all of the above criteria in #5, you are probably ready to go. Revise your question if necessary based upon what you now know.
IV- Developing a Hypothesis
The process of forming a clear hypothesis and devising acceptable tests is called "strong inference". It is described in a famous paper by Platt. Click here to read: Strong Inference. In the paper, Platt also describes why you should devise alternative hypotheses.
Try developing several questions related to your topic, and then choose one that best fits. Rephrase your question until it is both analytical, requiring analysis of data, and focused and based upon previous information on the topic to prevent a superficial discussion. Now try to rephrase your "best fit" question into a predictive statement that clearly defines the variables under investigation. Some examples follow:
- Arthropod communities of old growth forests will have greater species richness than those of clear cut forests.
- Areas dominated by non-native plant species support fewer bird species than areas dominated by native plants.
- Roots of pea plants treated with Rhizobia will have greater biomass than untreated roots.