Writing reports

Communicating your results

  • Being a statistician means being able to do several things:

    1. Obtain and process the data for analysis
    2. Do a suitable analysis
    3. Check that the analysis was reasonable
    4. Communicate your findings to the world
  • Last part is perhaps the most important: you do an analysis to answer a question, and the answer to the question is the most important thing.

  • This is true whether you are in the corporate world, answering to a boss, or in graduate school, where you will eventually have to convince your thesis committee (and, by extension, the academic world) that what you have done is interesting, statistically sound and important.

Reports

  • Final step of your process is to write a report. This is a sales job, because you have to convince your readers that what you have done is worth their time reading.
  • Writing a report requires good language skills. You cannot become a good statistician without that.
  • This is why so many of my questions end “explain briefly”. You need to learn to provide a complete and concise explanation of what your results tell you and why.
  • Reports are usually structured in a similar way, as shown on next page.

Report structure

  • Introduction: tell your readers about your problem and what you hope to find out. Provide enough explanation for the reader to know what you’re trying to achieve. Can also refer to what other people have done.

  • Methods: Where the data came from, how collected (describing technology used, if any). Scientific people call this “Methods”. Also here: describe work to get data into right form.

  • Analysis and results: Not enough to give analysis; have to explain what you are doing and what made you do it. Describe results in matter-of-fact way (opinions in the next section).

  • Conclusions: What does analysis tell you about your problem? Place results in context. Offer (supported) opinions about what the results mean, to you and the world.

A typical journal article

Link here.

Title and authors, with journal and page numbers, so that you have everything you need to refer to it.

Abstract

Journal articles typically begin with Abstract that summarizes question and gives highlights of results and conclusion, and tells you whether paper is worth your while to read.

Introduction

Introduction begins with plain-English first sentence. The numbers in brackets are references to what other people have said.

Materials and methods

The subjects. Experiments on humans require “ethical approval”.

Taking measurements

…and

Results (a)

…noting that the two groups were not significantly different before the study, but changed in important respects over time. Results also shown in table.

Results (b)

Graph showing that bone mass density has changed greatly as a result of the jumping. (Graphs are always good.)

Conclusions (selected) 1/2

Conclusions 2/2

Note use of (relatively) plain English, description of most important findings, comparisons to other work, and admission of limitations.

References to other work (some)