Watch the Video
Watch 5 Common Stats Questions from Early Career Researcher (approx. 6 minutes long).
Test Your Knowledge
You’re planning to do a study. When is the best time to talk to a statistician?
A. Before you determine your hypothesis.
B. When you’re designing your study.
C. After you’ve collected your data.
D. Once you’re trying to choose what formula to use for analysis.
E. Never. Statisticians are in a separate field and won’t be able to advise cancer researchers.
The correct answer is B.
You don’t have to know all there is to know about statistics for cancer data science, as long as you recognize the importance of engaging statistical expertise as needed. Statisticians can provide advice on study design and choice of analysis methods or assist with analyses. You can contact NCI and other NIH statisticians who may consult on your research project or join the research team as full collaborators.
The incorrect answers are:
A. You should determine your hypothesis before reaching out to a statistician. Once you know what question you want to answer, the statistician can help advise you on study design.
C. You should talk to a statistician when designing your study. If there’s a flaw in your study design, it may be too late to fix by the time you collect and start to analyze your data, and your findings may not be statistically valid.
D. You should talk to a statistician when designing your study. If there’s a flaw in your study design, it may be too late to fix by the time you collect and start to analyze your data, and your findings may not be statistically valid.
E. Statisticians are able (and often happy) to advise cancer researchers. They can help you avoid serious flaws in your study design.
Related Materials
- Reporting Guidelines: Find reporting guidelines for a wide variety of health research studies under EQUATOR (Enhancing the QUAlity and Transparency Of health Research). By looking at reporting guidelines relevant your type of study and their accompanying explanatory publications as you are planning your study, you’ll have a good sense for what you should anticipate needing to report about your study and why all of those aspects matter for others to evaluate the quality of your study and properly interpret its results.
These courses require sign-up to attend. Keep an eye on their calendars to see when they become available.
- Statistical Inference for Non-Statisticians: Part 1: You’ll learn the basic thinking behind two schools of statistical inference.
- Overview of Common Statistical Tests: Part 1: Provided through the NIH Library, this course covers the general concepts behind statistical tests.
- Overview of Common Statistical Tests: Part 2: You can attend both, or either, and still gain valuable understanding of how to understand and prepare data, interpret results and findings, design and prepare studies, and understand results.
- Overview of Common Statistical Tests: Part 3: This segment describes basic concepts for using common statistical tests such as Chi-square, paired and two-sample t-tests, and more.
- A Review of Epidemiology Concepts and Statistics: Part 4: This session gives you the opportunity to explore statistics for epidemiology.
Keep Going
Continue to Chapter 4 to learn about big data technologies we think can accelerate your education and research.
- Ready to start your project? Get an overview of the data science lifecycle and what you should do in each stage.
- Need answers to data science questions? Visit our Training Guide Library.
Instructor
Lisa McShane, Ph.D., NCI Division of Cancer Treatment and Diagnosis (DCTD)
Dr. McShane is the associate director of NCI DCTD’s Biometric Research Program. She is an internationally recognized expert on precision medicine clinical trial design, the development of tumor markers and omics predictors, and reporting guidelines for health research studies.
For questions and feedback about this chapter, email our team at ncicbiit@mail.nih.gov.