The Joy of Stats

In previous years, I've taught a course in statistics and data analysis to final year biology undergraduates. With this comes a sort of weird affection for all things statistical. It's not the coolest thing in the world to admit to, but i think the cool boat has long sailed for this particular scientist!

Last nights "The Joy of Stats" by the ever-enthusiastic Prof. Hans Rosling on BBC4 was a tour de force in public service broadcasting and science communication. Rosling's obvious enthusiasm, combined with state of the art graphics and excellent contributions from other leading statisticians made the whole topic come alive.

Unfortunately, BBC iPlayer doesn't allow you watch the show online if you are outside the UK, but it will be repeated on Monday next at 11pm on BBC4. In the meantime, you can enjoy this wonderful clip from the show where Rosling demonstrates how you can condense huge amounts of data on public health and wealth and make it interesting and exciting!

Applaud your ANOVA...it's World Statistics Day

It's time to celebrate your null hypothesis, to thank your t-test and applaud your ANOVA; because the United Nations has dedicated this day to celebrating all that's great about 'official' statistics. Today, 20th October 2010 is the first ever World Statistics Day!

Indeed, Ban Ki-Moon, UN Secretary-General, has welcomed todays celebrations saying that statistics "permeate modern life" and that they affect all our lives:

"They are the basis for many governmental, business and community decisions. They provide information and insight about the trends and forces that affect our lives. Collected in surveys and censuses - three billion people will participate in population and housing censuses this year alone."

Of course, what the UN had in mind with WSD, I think, was the big statistics. That's the statistics collected on the grand scale by governments in censuses of populations.

My own interest in statistics (and most scientists, I think) stems from the statistics collected in our own individual experiments. So, we want to see whether one set of 30 plants are significantly different in height than another set of 30.

A scientist might want to know whether the concentration of a particular pollutant is higher or lower in two seperate lakes. So, she will take a defined number of samples from each lake and compare.

The important thing about such experimental statistics is that we are not sampling all the water in the lake or all the plants in the world. We are taking a defined number of (often) random samples which we hope will reflect the overall spread of the variable (be it height or pollutant concentration) throughout the entire population.

Nevertheless, it's pleasing to note that the spotlight is being turned on statistics for once. It's not a sexy subject. Having taught statistics to biologists for a number of years, I am familiar with the groans that accompany any mention of the term statistics.

However, a good understanding of statistics and experimental design is essential for all scientists and an understanding of 'good' and 'bad' statistics is useful for everyone to allow us to weed out the large amounts of erroneous "statistics" that pop up in the media from time to time.

No, statistics may not be sexy, but they are very very useful.

Happy World Statistics Day!