What statistical technique will be most valuable in my work?

Question from a student (paraphrased): "Is there a statistical technique that will be most valuable to me in my work?"

Answer: There is no such technique. There is  no panaceas in public health statistics, as there is  no cure-all for social ills. The best advice I can provide is to gullibility and to follow common sense and "the scientific method" whenever possible.

Question: "What is the scientific method?"

Answer: I cannot answer this question directly. The best I can do is to illustrate the benefit of the method and hope that you will take it as a matter of faith that common sense and the scientific method are the best "cure" for our ills and societal problems. 

Consider the intellectually curious haberdasher who first studied population health on a scientific basis. His name was John Graunt, and he lived from 1620 to 1674. Epidemiologists, demographers, and insurance adjusters cite him as their  hero. Graunt's most famous work is Natural and political observations ... upon the Bills of Mortality ). You can read this document yourself by clicking on the highlighted title. There is nothing like reading a primary source. 

Graunt's method has contributed to our well-being. A look at Graunt's survival table from 17th century England quantifies the benefit: Of 100 people born in 1600, 64 lived to age 6,  40 to age 16 years, 25 to age 26 years. Without the Enlightenment, most of you would now be dead. We are lucky men like Graunt existed.

I still have not answered your question directly. Context is important. There are diverse views as to what makes a science, but three constituents will be judged most essential ( John Tukey 1915-2000, 1962, p. 5):

  (1) Intellectual content
  (2) Organization of data into an understandable form
  (3) Reliance on tests of experience as the ultimate standard of validity.

Point (1) requires us to pay attention to data quality, study design, and other esoteric components of a study. There is no getting around this. Do data measured what they purport to? How good are measurements? Is the sample adequate in quality and quantity? What other variables enter into the relation being studied? Validity is critical.

Point (2) recommends use of plots and summary statistics. Without understandable statistics, we have little context. Always start analyses with simple to understand numerical summaries and plots.

Point (3) raises questions about induction and the status of evidence. What evidence counts and what evidence can be discounted? Good practice places emphasis on reasoned judgment based on accumulated evidence, and not on dogmatic insistence of the unique validity of a certain technique or way of thinking. 

Avoidance of fads about science being a social construct need not undermine the  present state of knowledge. We can accept that science is a social construct and that the current state of knowledge is uncertain while  recognizing the difference between scientific knowledge and belief. "An important proof for the value of this process is that science works, from its ancient forms in the construction of temples or cathedrals to telecommunication and experimental gene therapy" (Vandenbroucke & de Craen, 2001, p. 512).

Portada de Las Observaciones