I recently met with a student working on their first major project.
The student had ticked the boxes on the empirical script.
Interesting question - check!
Unique data - check!
Rigorous method - check!
Yet.
Months into the analysis.
The paper was not done.
Why?
Because the student was caught in an endless loop of data exploration.
The research team could not agree on an approach to data analysis, a set of hypotheses, and a general storyline.
Being a junior team member.
The student was asked to run and re-run data.
And felt they could not say no.
It’s a terrible place to be.
More weeks into the project.
They were still running data.
Eventually.
I suspect the student will give up on the project.
It does not have to be this way.
Even if junior.
A #PhDstudent can shape a project.
How?
First, before ever crunching data, have a clearly defined #researchquestion or set of objectives.
This gives you a roadmap for what analysis is to be done.
Second, stick to your roadmap.
You can waste a lot of time exploring side questions.
Finish the original project.
You can answer side questions in the next paper.
Third, come to meetings prepared.
The analysis you present and the confidence with which you answer questions will shape what you are asked to do next.
Have the analysis formatted in an accessible way.
This evokes confidence in your work.
Fourth, don’t open the door to more data exploration.
Nothing is more frustrating than a student saying ‘it didn’t work so I tried …’
But then not presenting the complete analysis - or saying they will the next time.
This results in asks for even more analysis.
Fifth, document what you have done.
Often analysis doesn’t work out.
Document it.
That way, you don’t have to do it again or you can explain why another round of analysis doesn’t help.
Sixth, stop chasing the method of the month.
After conferences, students always say why don’t we try xyz that I just saw.
That’s great - except it usually doesn’t fit the question or data.
Pick a method that fits the design and stick with it.
Let the #reviewers decide if you need to do something different.
Finally, communicate.
Advisors often don’t understand the time it takes to learn and run an analysis.
Ask for help if you need it & explain the time required.
Often, this will result in helpful comments and more modest requests.
By being prepared, staying focused, and communicating, you are much more likely to emerge from the fog of data analysis with a completed paper.
Best of luck!
Comments