In the last ~5 months or so in my gap year position, I’ve come to realize how statistics can potentially be important for pre-meds interested in research. Specifically, clinical research.
I helped teach introductory statistics at Cornell for 5 semesters, to a mix of non-majors. Of course, the majority of them are not pre-meds, but I’ve also taught plenty of them during my time as a TA. And it’s always a similar story: they took that course not because they want to learn statistics, but because they want an easy A.
I’m not saying you have to really truly enjoy statistics, but I think there is truly value to learning the subject beyond just the basics (intro stats, intro biostats, and the like), and looking beyond the “easiest class to take so I can get an A”.
The benefits of understanding more than just T-tests
You’ll understand papers more easily
You’ll know exactly what the “methods” portion of a clinical research paper is talking about, and furthermore, having a solid understanding of statistics will allow you to look up unfamiliar models/analysis and get a grasp of it quickly.
For example, if a paper I’m reading uses Group Based Trajectory Analyses, I might look that up and see that it’s an extension of linear mixed models–and because I know what LMMs are, I also get a intuitive understanding of GBTMs in the paper.
You’ll understand that p-values aren’t the “end all, be all”
This really frustrates me as everyone references p < 0.05 as “significant” and aim for that value. P-hacking is rampant, biostatistians are asked to fake data, and professors have also had to get their papers retracted because of poor ethical choices.
But actually, whether or not you’re able to detect this significance can be based on your sample size. In my freshman year, my first ever statistics professor said, “if your sample size is large enough, anything can be significant”. There are, of course, cases where that statement is not true, but the general idea is that the larger the sample size, the smaller the effect size needs to be to detect a difference.
It is merely a convention, and while it’s a valuable tool for us to understand our data, we shouldn’t rely on a single metric and try to manipulate things so that we get that p < 0.05.
You have an intuitive understanding of what kind of analysis YOU might need
A T-test or a Rank-sum test?
Of course that is probably the simplest question; but of course if you intend on one day running your own lab (and even instructing your own RAs!), you need to intuitively know what kind of analysis you should run, its assumptions, caveats, and interpretations.
The statistics I think pre-meds should know
- What is considered to be “intro”–the basic t-tests, ANOVA, linear regression
- Categorical Data Analysis–Exact tests, Chi square tests, agreement statistics, logistic regression, ordinal logistic regression, poisson regression
- Survival Analysis (Kaplan Meier, Cox Proportional Hazards Models)
- Model fitting and evaluation (Goodness of fit tests, Likelihood ratio/Wald tests)
- Mixed models (growth curve analysis/nested models)
- CODING! Know some R, Python, & basic machine learning beyond Logistic Regressions–this will serve you incredibly well!
All of this, and their assumptions, downfalls, and interpretations.
Of course, it’s not a comprehensive list, just a few things I’ve encountered so far in my journey in clinical research and analyses that come up again and again.
Final Thoughts
Take some time taking 1 or 2 additional statistics classes than necessary. Coding (in R, Python) can be extremely beneficial too data wrangling is just as big of a part as analysis.
Statistics has been a unique part of my pre-med journey, and while I know not everyone may have my passion for it, it’s certainly a very useful skill for any pre-med interested in clinical research!
Anything I missed? Any further thoughts you have? Or do you simply disagree with me–please let me know your thoughts down below.
