It is easy to see. Every day, I do it, you do it. We all have an attraction to the latest and greatest: the new iPhone, a pair of shoes… or for this crowd, the latest Illumina array or hyper dimensional causal model. We feel immediate wonder when introduced to a new technology or technique. That wonder can drive us to create, but if we are not careful, it can blind us to our goal. In public health we have but one goal, that is to increase the health and wellness of all people. That goal is not to do science;
it is to use science to combat disease.
Ifind this important because as a young student of epidemiology I find myself attracted by novelty. I “oooh” and “ahhh” at complex analyses, and yawn at simple studies; yet novelty should have no bearing, as our work is for the greater good. I am by no means a Luddite—quite the opposite—but therein lies the temptation.
New tools are sexy. They get the most press, are more likely to be published in Science, and, no one can deny that doing something “new” is almost always more exciting. But we cannot let ourselves fall into the trap of allowing technology or method to dictate our questions. That is, we cannot allow our hypothesis to be driven by the tool; our hypothesis must be guided by our goal. In public health research, that goal can seem a bit abstract, but the principle is directed by the thoughtful plodding that is hypothesis testing. We have heard it all before, but there is wisdom in the scientific method that transcends your study and mine. We are all (or should be) hypothesis testing.
That is, formulating a sound hypothesis and systematically collecting and examining data using any tools necessary in order to confirm or deny.
Yet fundamentally we first have to pose a question, and this can be the most difficult and important part of the entire process.
Over the last two years I have had the privilege of working closely with Dr. Warren Winkelstein Jr. During our studies into the history of epidemiology, an interesting yet obvious pattern has emerged. Many of the great advances in public health were not the result of a technology finding a question; they were the result of a hypothesis finding a tool. “Rubbish,” you say, but there is evidence:
- In the 1760s George Baker did not find a disease to apply a test to. He used possibly the first public health laboratory to test for lead to answer his hypothesis on the Devonshire Colic.
- During the 1850s John Snow did little more than count, but in the process of systematically testing his hypothesis on cholera, he founded epidemiology.
- In 1912 Janet Elizabeth Lane-Claypon did not look for two groups on which to use a t-test; she needed a way to compare groups of women to complete her seminal paper on breast cancer.
It might sound silly to think of searching for a question on which to use a t-test, but imagine what researchers in the future will think of our tools. When viewed in the long term, the wisdom of hypothesis testing shines. You may know your tool well; but if that tool is a hammer, you look for nails, and you will never see the house. As researchers we must keep an open mind. We must place our effort into posing important questions and then find the tools needed to test our hypotheses.
The elegant study that describes what others have failed to see is the way we advance in science, yet it is our responsibility to ensure that public health research attains a deliberate and careful equilibrium. How do we effectively balance the use of dazzling, novel tools with hypothesis testing that maximizes public health impact? This is the student’s dilemma.
Steve Francis is currently a doctoral student at the School of Public Health. He is a graduate student researcher for the Northern California Childhood Leukemia Study.