Just this past September 2020, we (PhysioQ https://physioq.org/stem ) worked with CustomED (another non-profit, https://customed.org/ ) to run the Data for a Difference nation-wide competition pilot, where students across the USA aged 13-18 used our LabFront For Schools platform and wearable wristband devices to collect and analyze data around a health problem in their communities over a 2-week period. Each team was partnered with an experienced researcher who served as mentors throughout the whole competition. I happened to mentor three very bright, motivated students from Wisconsin.
At the very start, the students were trying to figure out which research topic/question to tackle for the short 2-week competition. To help them out, I sent them this email:
______________________________________________________________________________________
Team, I look forward to hearing your ideas/thoughts tomorrow!
Ideally, by the end of tomorrow, we should have narrowed down the research question, hypothesis, protocol and analyses – considering that data collection starts Monday. I think we should also leave some room for maneuverability too – just in case the protocol doesn’t work as planned and we can make adjustments as necessary.
As a researcher, I have found out that it is MUCH harder to find the right question to answer than finding the right answer to a question. Many researchers follow a script and do research without thinking of the big picture, and as new scientists, I encourage you not to do this: think big first, before diving into the details. The reason finding the right question is hard is because you need to sometimes intuitively feel where the gold-mines are before actually getting the map. More specifically, you need to :
-
know which research question will have the greatest impact (be important, relevant, and ideally actionable)
-
yet recognize what is doable (i.e., scientifically rigorous and methodologically sound) with the tools (metrics, analytical, time/physical resources) that you have at your disposal
-
feel where there is greatest chance of finding the most intriguing results
-
while sensing what really arouses your curiosity – are you excited to see the results and find out what you discovered?
Your research question should ideally meet these criteria. Of course, given the short time period and constraints of the competition, we can’t be overly ambitious.
Personally, what I do is write out a number of questions to ask myself: Why do you feel this research question is important? How will it change people’s behavior, perceptions? Are the data going to be good enough to address the question? Why am I excited to study this? I jot outline answers to these questions – and this process gives me kernels of ideas to put together and figure out which question is the best to pursue. On the topic of social media there are so many options (as we had discussed): why is addictive
After you have gotten the question down, then I write out the Objective, Specific Aims, and Hypothesis.
The Objective provides an overarching aim of your study. In our case, it may be pretty straightforward: “Understanding the role of Social Media on Mental Health (or productivity?)”
Specific aim then narrows down the objective into a more detailed question that can be addressed in one study. For instance: “To determine the physiological mechanisms behind the association between asthma and seasonality” or “To assess whether eating out more is associated with greater weight gain”
Hypothesis then forces you to take a stand on this question – a stand that should be supported or refuted by your experiment. An often times, there can be multiple hypotheses under a specific aim:
“Higher amount of pollen in the spring time is associated with greater incidences of asthma presentation to the hospital”; or “Woodfire stove pollution during the winter is not associated with increased asthma medication use”.
As you can see, this is where you can define further what you mean by “Mental Health”: stress levels as determined by Garmin watch? lack of sleep? higher heart rate? etc…
These hypotheses will then assist you on figuring what your protocol and metrics are going to be….
I hope this all helpful. Happy to address any questions on-line too before our meeting tomorrow. Happy researching!
-Andrew