Genomics Blog

January 19, 2009 2:00 PM
The Drunkard’s Walk – a great review of statistics.
Filed Under: Gerry Ward

A couple of Biology curricula ago, there was a small segment – one learning experience – devoted to probability. This particular lesson involved students learning the difference between the inferred ratios and observed ratios. Clearly this is an area that should allow and encourage cross-curricula study.








One trouble in modern high schools is that lessons are presented episodically and not even necessarily in anymore logical fashion than putting pieces of a jig-saw puzzle together. It takes many pieces to be put in place before the picture starts to become clear. The other issue that creates difficulty for students is the misleading use of numbers and mathematics repeated often in the media. Take for instance the recent commercial on television for a type of wiper blade sold at a major Canadian tire store. The woman proclaims that she doesn’t listen to weather reports anymore because they said there was a 30% chance of snow….and it snowed. She announces proudly, ‘they were wrong’. This commercial is repeated probably a dozen times per day or more. A student watching television would soon start to think that a 30% chance of something happening means that it won’t. Of course, this is rubbish!

There are many ways in a classroom of generating data for student analysis. For example, I have seen computer programs instantly generate the kind data that would be obtained in hundreds of generations of pea plants to simulate numbers related to Mendel’s experiments. I’m not sure that students really grasp numbers generated in such a manner. I preferred the good old fashioned coin flipping. You know, the kind where students work in pairs. Each has a coin with the alleles marked in masking tape so that a flip is like the equivalent of a meiotic division. Then when the coins land, it is the equivalent of fertilization. In this manner, the students actually experience seeing the numbers generated and appreciate the independent manner each flip is. The students can readily start to see that the observed ratios approach the inferred ratios only as the number of flips increases. This is a good time to bring into the discussion the idea that there are ways of deciding when the differences between observed and inferred is a real difference, or only one that is due to chance. It is a good opportunity to tell students that scientists don’t just look at the data and proclaim “that’s close enough!” Scientists actually have statistical methods of analysing the data. Some students will even want to give that a try themselves, especially if they have already been exposed to the methods in their mathematics class.

I recently read Leonard Mlodinow’s book “The Drunkard’s Walk: How Randomness Rules Our Lives”. I found the book a fascinating array of mathematics, history and anecdotes. I especially like the way Mlodinow presents the history. He talked about the scientists and what things inspired them to come up with their ideas. It is probably not surprising that the earliest developments in the study of probability came from gamblers and games of chance. This book gives teachers many anecdotes to illustrate probability and statistics in an understandable way.

Although the book does not mention the word gene, genome or genomics, it does discuss some of the mathematical reasoning related to DNA evidence used in court cases. Teachers will also be amused (I hope) with his analysis of student evaluations. He points out how extremely difficult it is to remain objective and come up with a reliable and valid mark for student work.


In all aspects of science education including Genomics, an ability to understand the meaning of the statistics is important. I highly recommend taking in this easy read.


Comments

Bev Jaremko - beverleyjaremko.webs.com

As a teacher I completely agree. Mr. Ward's comments are bang on. Too often when I am teaching kids' eyes glaze over when we talk of math at all and it is vital to make the lessons visible, concrete and if possible fun to demonstrate the concepts. I know of one teacher who uses card games to teach probability- he has a PhD in math but bless his heart he gets right down and concrete to teach kids. I have heard a criticism of our education system in general that we piecemeal the information and just repeat it every few years hoping kids get each exposure a bit more understanding of fractions, decimals, algebra. What I did with my own four kids was a gentle program to walk them through math in stages of logic and there was a plan and a progression. I did the same with science, and with history. I don't know why the curriculum does not. Tot each history and science chronologically is eminently logical because that is the way it happened. When little kids are taught insights of each period of time they don't even have the bias of confusion from having heard the ideas before. They also learn the ideas chronologically, the way humankind did. So when I was for instance teaching the kids about the Archimedes screw or the discovery of a new element for the periodic table or the invention of the zipper, they knew what background the scientist had to work with when he had this insight. When kids 'think back' to earlier times, they actually are 'thinking back'. It's quite fun. I also have to add that in terms of teaching math, logic, philosophy and statistics a very wise teacher colleague shared with me her insight once that we do it exactly backwards. We leave philosphy and logic till senior high but teach math young. In fact kids understand very early about equality of cookie distribution and about fairness and sharing, and they understand very early on a philosophical level about depth of snow and size of boots needed. We can add the math later. Maybe kids would enjoy and use math more if we taught it more like Mr.

RW -

This book does a decent job of explaining the important points of statistics and probability.

Statistics courses were highly misunderstood and underutilized by teachers, but there is likely good reason for it. They were hard to understand and easily forgettable because they were poorly taught at university and not only in the education faculty.

Having taken a couple of those "introductory" statistics courses—including the education faculty introductory statistics course which was oversimplistic and likely still is a joke —I remember the difficulty of the "real" stats courses and remember being dismayed at the high drop-out rate of students from the courses I was in (and successfully completed).

As an aside, one must keep in mind that few university types and even fewer math university types know how to teach. They know how to lecture; they don't have a clue about teaching

I admit I struggled with the courses and I have to say that Mlodinow's book, while absolutely entertaining, is not for the faint of heart would have been helpful. The book might require one to sit down and read parts more than once— perhaps even several times—to grasp what he has to say.

The book is an excellent read. Just be prepared to take time to understand it to follow the logic. It's even easier than an education stats course if you do so.

Name
URL (remove the http://)
Email
Comments