Post Conference Goodies

Photos

Photos are available on Facebook (click here) and Flickr (click here). Note, you’ll need to request to join the Facebook group.

Presentations

Should you not want your slides or poster made available, please inform Vanessa Cave (vanessa.cave@agresearch.co.nz).

Consulting in the real world: Communicating statistics to scientists

During this special session, there were four invited talks on statistical consultancy followed by a discussion. To assist the discussion, evpoll was used to capture anonymous comments (with voting on the comments). These comments are given below.

“Moving beyond p”: How can we best promote awareness around the aim of a statistical analysis?

Comments	Up votes	Down votes
Data patterns, distributions, confidence intervals	12	0
Help people understand what their research question actually is.	12	1
Get “statistical thinking” embedded rather than just what’s the answer	10	0
Focus on practical significance	10	0
Providing courses in statistics that are delivered by staff from the statistics school as opposed to staff in other faculties with both minimal understanding and enthusiasm for the subject	5	0
Massively improve how we teach stats in the first place	6	1
How do we make non-significant results publishable?	9	5
Change thinking from how to get results to what we want to know.	2	0
Relate output back to how it addresses the aim	2	0
Simply ask the question first: can you see the effect on the graph…	2	0
What about Bayesian approaches?	3	1
Convince editors that in some cases a picture is far more informative than any p-value or complicated stats.	0	0
Don’t provide p-values 😁	2	6

Basic Statistical ideas: Ways to encourage sound understanding ideas/ philosophies (e.g. replication vs pseudo replication)?

Comments	Up votes	Down votes
Training using the researchers examples	15	0
Develop a virtual reality maze with statistical tests along the way and only talk to them when they find their way out 😉	6	0
Experimental games that users participate in and see variability between their results and others	8	2
A chocolate bar for thinking about design for just one minute	7	1
Use computing simulations	6	1
Get involved in the experiment running stage, going to the lab, that way you can better understand critical considerations they have, and address these.	6	2
We should go to none stats conferences and vice versa	3	0
Ensure the basic concept of variability and “effect vs noise” is understood	3	0
Support concepts using practical and relevant examples	4	1
Explain with examples	3	0
Use fun, engaging examples!	3	0
Show striking counter examples	2	0
Graphical interpretation of interaction effects.	1	0
Teach school kids statistical ideas like this – they will get it	2	1

Communicating results: Relating things back to the original question; Encouraging good standards in stat aspects of results presentation?

Comments	Up votes	Down votes
Statisticians should be involved in writing papers and producing the presentations	20	0
Encourage people to think about what the results are showing and do they make sense. Is there a scientific explanation.	12	0
Support good writing of statistical results and conclusions	9	0
Have the original hypothesis explicitly written down and ask how does this answer that	7	0
Publish research/analysis protocols before data collected (cf Clinical Trials)	5	0
discuss the limitations, why and where it wont work	5	0
Reporting guidelines used in the design phase eg consort statement	2	0
Use graphics.	2	0
Create graphs with captions for the sci	2	0
use layman terms	1	0
Give examples!good results vs bad result	1	0
Teaching that to students in first year courses.	1	0
Culpability	1	0