Integrated Primary & Secondary Research

6 Process of Conducting Secondary Research

A syringe with pink fluid going into one test tube in a box filled with test tubes.
Photo by Louis Reed on Unsplash

Step 1: Define your research topic and question

  1. Start with a thorough literature review
  2. Ensure that the research question has clinical or policy relevance and is based on sound a priori reasoning. A good question is what makes a study good, not a large sample size
  3. Be flexible to adapt your question to the strengths and limitations of the potential datasets

Step 2: Select a dataset

  1. Use a resource such as the Society of General Internal Medicine’s Online Compendium
  2. To increase the novelty of your work, consider selecting a dataset that has not been widely used in your field or link datasets together to gain a fresh perspective
  3. Factor in the complexity of the dataset
  4. Factor in dataset cost and time to acquire the actual dataset
  5. Consider selecting a dataset your mentor has used previously

Step 3: Get to know your dataset

  1. Learn the answers to the following questions:
    • Why does the database exist?
    • Who reports the data?
    • What are the incentives for accurate reporting?
    • How are the data audited, if at all?
    • Can you link your dataset to other large datasets?
  2. Read everything you can about the database
  3. Check to see if your measures have been validated against other sources
  4. Get a close feel for the data by analyzing it yourself or closely reviewing outputs if someone else is doing the programming

Step 4: Structure your analysis and presentation of findings in a way that is clinically meaningful

  1. Think carefully about the clinical implications of your findings
  2. Be cautious when interpreting statistical significance (i.e., p-values). Large sample sizes can yield associations that are highly statistically significant but not clinically meaningful
  3. Consult with a statistician for complex datasets and analyses
  4. Think carefully about how you portray the data. A nice figure sometimes tells the story better than rows of data


This page contains materials taken from:

Smith, A.K., Ayanian, J.Z., Covinsky, K.E. et al. Conducting High-Value Secondary Dataset Analysis: An Introductory Guide and Resources. J GEN INTERN MED 26, 920–929 (2011).


Icon for the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

An Open Guide to Integrated Marketing Communications (IMC) Copyright © by Andrea Niosi and KPU Marketing 4201 Class of Summer 2020 is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, except where otherwise noted.

Share This Book