From this tweet: click on the image there to enlarge
The 10 most-visited posts at Statistically Funny:
1. What's So Good About "Early", Anyway?
In which a reluctant early bird questions a cliche. A clue to what you need to know about screening tests.
2. Nervously Approaching Significance
A 2013 post about statistical significance. A better starting point now is my 2016 Absolutely Maybe post, 5 Tips for Avoiding P-Value Potholes.
3. ARR OR NNT? What's Your Number Needed to Confuse?
Love the numbers: but it doesn't mean they're entirely objective, either. On words and misleading numbers.
4. The Forest Plot Trilogy - A Gripping Thriller Concludes
Mysteries of meta-analysis plots revealed.
5. Data Bingo! Oh No!
Why do so many research findings turn out to be rubbish? Here's one reason: the perils of multiple testing explained.
6. Women and Children Overboard
Catch-22: Clinical Trials Edition.
7. If At First You Don't Succeed...
Significus the Obscure helps explain the ever-present dangers of post hoc analyses.
8. You Will Meet Too Much False Precision
A fortune teller shows us the future: where everyone expects to know the confidence intervals.
9. Does It Work? Beware of the Too-Simple Answer
Breaking questions right down. In which we go statistically Bayesian, too.
10. Biomarkers Unlimited: Accept Only OUR Substitutes
Surrogate or intermediate endpoints and biomarkers: know one when you see one! (And you'll see a lot of them.)
The runners up tackle:
- Composite endpoints
- Standard deviations
- Preclinical animal studies
- Statistical literacy
- Absence of evidence doesn't necessarily mean evidence of absence
As of 3 September 2016
No comments:
Post a Comment