01 Statistical Analysis
Here is the updated presentation for 2009, with more information on Excel and a worked set of examples with hummingbirds, to tie in with the natural selection topics. Also, skip on over to the Excel StatBook resource, for a set of examples, tables, graphs and significance tests that you can play with.
And Geoff Browne kindly gave permission to upload his t-test powerpoint to slideshare:
Test your reaction times here! Are there significant differences between Friday before lunch and Monday after lunch?
Does the internet help us learn facts? Work in groups, design 10-question fact-based quiz. Give out, participants can’t use the internet. Collect scores, means, standard deviations. Give time to look up answers on internet. Test again the following day. Compare means, standard deviations, plot with error bars. Discuss results & method validity. You could go a step further and perform a ‘meta-analysis’ based on the results of all the groups combined.
Use the IBDP Bio Excel StatBook to choose, apply and interpret the right test and graphs.
Excellent Handbook of Biological Statistics from John MacDonald
Basic Statistical Tools, from the Natural Resources Management Department
And The Little Handbook of Statistical Practice is very useful.
Sumanas statistics animations
Field Studies Council stats page, including the t-test
And this enlightening talk from Han Rosling: No More Boring Data!
Click here for a funny article on the 9 circles of scientific hell.
Also, play with this: Google Correlate.
Using your calculator:
- Using the TI GDC (from Click4Biology)
- Using the Casio pdf download (from keymath.com)
Statistics in Action:
‘Real’ acupuncture no more effective than fake acupuncture, from ScienceDaily
Evidence Based Medicine First, medical website explaining the false health claims of many alternative medicines.
Here’s a nice profile on Edzard Ernst, the world’s first professor of alternative medicine. He has spent his career trying to get alt-med in line with real science.
Ed Yong, MrT’s blogging hero, writes for Cancer Research UK on the WHO’s verdict on mobile phones and cancer. Correlation vs cause!
Epidemiology: The Science of Cohort Studies. How do we generate lifetimes’ worth of data in studies in medicine? Ben Goldacre’s BBC Radio 4 documentary, Science: From Cradle to Grave. An amazing discipline to work in, and one birth cohort study has been running for over 65 years!
“Facebook Gives You Cancer” err…
And another peach from XKCD:
Key terms: t-test, mean, variability, data, reliable, significance, sample, excel, calculate, correlation, graph