In this episode, I explain how to come up with a good question, obtain the background information you need, find research, obtain full texts, organize them, read the different sections of a paper to get the right kind of value out of it, and critically analyze the study design. If you're a beginner, this is really designed for you. If you're more advanced, you'll enjoy the specific examples I give of problems interpreting research studies.
Listen on ITunes or Stitcher.
Click here to stream.
Right-click (control-click on the Mac) here and choose “save as” (“save link as” on Mac) to download.
Subscribe in your own reader using this RSS feed.
Want transcripts? Sign up for the CMJ Masterpass with this special link to get 10% off.
This episode is brought to you by US Wellness Meats. I use their liverwurst as a convenient way to make a sustainable habit of eating a diversity of organ meats. They also have a milder braunschweiger and an even milder head cheese that gives you similar benefits, as well as a wide array of other meat products, all from animals raised on pasture. Head to grasslandbeef.com and enter promo code “Chris” at checkout to get a 15% discount on any order that is at least 7 pounds and is at least $75 after applying the discount but under 40 pounds (it can be 39.99 lbs, but not 40). You can use this discount code not once, but twice!
“How to Read a Science Paper” Show Notes
In this episode, you will find all of the following and more:
0:06:15 How to develop a good question
0:09:30 How to use Pubmed and Google Scholar
0:11:50 Why and how to use MeSH terms (medical subject headings)
0:16:50 5 ways to get full-texts for free that are totally legal
0:24:35 How Sci-Hub will facilitate the technological evolution of research distribution and the Spotify-ication of the science publication industry
0:32:45 How to organize science papers to prevent wasted time and frustration later
0:34:40 Reference management software
0:36:35 The anatomy of a science paper; how you should approach each section and what you can learn from it
0:46:45 Peer review makes discussions within papers more objective; how a scathing peer review from six years ago continues to influence how I teach hormesis today
0:55:30 Acquiring background information with textbooks
0:57:35 Specific textbook recommendations
1:05:15 What you need to do before developing your own point of view
1:10:30 Strengths and limitations of different study designs
1:13:47 Observational versus experimental studies and the tradeoffs of context, size, and duration with strength of cause-and-effect inferences
1:16:50 The central role of randomization in experimental studies
1:19:20 Randomization needs a high sample size to be effective
1:21:07 Example: Finnish Mental Hospital Study
1:22:50 Example: LA Veterans Administration Hospital Study
1:25:50 Regression to the mean; how a study can show something to be true when it’s completely false; change-from-baseline data versus differences-between-groups data
1:35:45 The need for a control group: Atkins and methylglyoxal study as an example
1:37:35 Compared to what? Picking the right control group
1:41:50 The generalizability tradeoff: in vitro and in vivo, animal and human, sex, race, and other population differences
1:46:47 Contextual patterns determine outcome
1:47:50 Thailand zinc/vitamin A study as an example of nutrient interactions.
Textbooks Mentioned in “How to Read a Science Paper”
Various Applied Statistics books (no specific recommendation)
Molecular Biology of the Cell (free version, 4th edition)
Molecular Biology of the Cell (most recent 6th edition)
Biochemistry by Ferrier, excellent starting point for biochemistry
Biochemistry by Berg, Tymoczko, and Stryer (free version, 5th edition)
Biochemistry by Berg, Tymoczko, Gatto, and Stryer (most recent 8th edition, 7-day free trial for Kindle)
Web Resources Mentioned in “How to Read a Science Paper”
EndNote (designed for someone who writes science papers)
Evernote (I use this for storing and organizing practically everything)
Science Mag: Whose Downloading Pirated Papers? Everyone