8
Hilda Bastian @hildabast NASW Sciencewriters 2013, Florida 2nd Nov http://www.slideshare.net/ HildaBastian/ (Link listing on last slide) Statistically Funny blogspot tips to protect yours from data-led error Click on cartoon for link to post Absolutely Maybe at Scientific American Click on image

Science Writers 2013: 6 tips to protect yourself from data-led error

Embed Size (px)

DESCRIPTION

Slides from my talk at the annual meeting of the National Association of Science Writers (NASW) in Gainesville, Florida on 2 November 2013

Citation preview

Page 1: Science Writers 2013: 6 tips to protect yourself from data-led error

Hilda Bastian @hildabastNASW Sciencewriters 2013, Florida 2nd Novhttp://www.slideshare.net/HildaBastian/(Link listing on last slide)

Statistically Funny blogspot

6 tips to protect yourselffrom data-led error

Click on cartoon for link to post

Absolutely Maybeat Scientific American

Click on image

Page 2: Science Writers 2013: 6 tips to protect yourself from data-led error

1. Remember data don’t speak for themselves

Think through the context of what happened, the methods of collection and method of analysis before analyzing the data

See also: Who’s saying it? Academic spin(click thumbnail)

Page 3: Science Writers 2013: 6 tips to protect yourself from data-led error

2. Avoidover-precision

Always think about confidence intervals/margins of error, matching use of numbers & language to data certainty (47.23% or just less than half?)

See also: Standard deviations (click thumbnail)

Click on cartoon for link to post

Page 4: Science Writers 2013: 6 tips to protect yourself from data-led error

3. Look at each outcome separately

The strength & quality of data vary within a study. And always look for all outcomes (especially adverse ones), not just the ones researchers point you to.

See also: Alleged effects include… (click thumbnail)

Click on cartoon for link to post

Page 5: Science Writers 2013: 6 tips to protect yourself from data-led error

4. Don’t mistake absence ofevidence for evidence of absence

Click on cartoonfor link to post

Page 6: Science Writers 2013: 6 tips to protect yourself from data-led error

5. Don’t lookat one studyin isolation

See also: He said, then she said… (click thumbnail)

Click on cartoon for link to post

Page 7: Science Writers 2013: 6 tips to protect yourself from data-led error

6. Be as critical of the data you want to believe as you are about what you disagree with

Click on cartoonfor link to post

The biggest bias we have to deal with is our own.

Page 8: Science Writers 2013: 6 tips to protect yourself from data-led error

Link listinghttp://statistically-funny.blogspot.com http://blogs.scientificamerican.com/absolutely-maybePromising = overhyped + under-tested:http://statistically-funny.blogspot.com/2012/06/promising-over-hyped-under-tested.html Academic spin: http://blogs.scientificamerican.com/absolutely-maybe/2013/09/09/academic-spin-how-to-dodge-weave-past-research-exaggeration/ You will meet too much false precision: http://statistically-funny.blogspot.com/2013/02/you-will-meet-too-much-false-precision.htmlStandard deviations: http://statistically-funny.blogspot.com/2013/04/dont-worry-its-just-standard-deviation.htmlStatistical significance: http://statistically-funny.blogspot.com/2013/03/nervously-approaching-significance.htmlAlleged effects include…: http://statistically-funny.blogspot.com/2013/07/alleged-effects-include-howling.htmlStudies of cave paintings have shown..: http://statistically-funny.blogspot.com/2013/06/studies-of-cave-paintings-have-shown.htmlStudies in isolation – “Look Ma, straight A’s!”: http://statistically-funny.blogspot.com/2013/04/look-ma-straight-as.htmlHe said, then she said…: http://statistically-funny.blogspot.com/2013/05/he-said-she-said-then-they-said.html Begging hopefully for less bias: http://statistically-funny.blogspot.com/2012/06/begging-hopefully-for-less-bias.html