Heads you learn, tails you learn.

Yesterday I posted a flashy way of teaching Mendelian genetics.  Today I want to post a super easy way of teaching basic Mendelian genetics. How easy?

Easy as flipping a coin.

Anyone who has ever taught genetics knows the most important word you will say during those lessons is “probability”.  If your students are like mine they will usually say something like “yeah, but it doesn’t mean it is definitely going to happen.”

And they are right. Sort of.

I will often say to them “You are correct, it doesn’t mean it is definitely going to happen in each single instance, but over time the numbers will show a larger pattern that falls into perfect probability prediction.”  Then like any good scientist, I prove it to them.

But not without luring them into a false sense of security.  I use a coin flipping lab to show them laws of probabilities and how with enough data the anomalies fade away into a beautiful perfect dataset.

First step is we flip a coin ten times.  This gets them excited because the numbers almost never come out 5:5 and they start telling me I was wrong. In fact in the three times I performed this coin flip I got….

9:1

7:3

4:6

These are nowhere near the 5:5 and if you know how to find percent error.  You will see these three ratios reveal a huge percent error.  Now this is where I give them the speech on how science is only as good as the data you collect. I explain the more numbers you have the less error you will have, if you are correct in your hypothesis that is. In fact I told them we will get our percent error under 1% during the class.

When each group was done they counted their heads and tails. Most were within 10 of the 50:50 exact value. Some dipped below that, but this is to be expected, because the more data we have the more the anomalies fade away.  So we added up all twelve groups and out of 1200 flips we had 607:593. A mere seven flips from perfect. This is a percent error of  .6% well under 1%

The smiles slowly begin to creep across their faces as they see the data come to life as facts.  Numbers actually being gathered relentlessly into a story of how probability isn’t just a maybe, but with enough data, a definite.  I repeated this with my next two classes and out of 3800 flips we got 1879:1921.  It comes out to a percent error of .5% little lower than before, and still, way under 1%

The students got so into it that the ones from the earlier classes kept coming in to look at the data going on the board to see how close we were getting.

However, I am still not satisfied. I am not sure what it is, but I love, LOVE doing this lab.  I think I have an unhealthy obsession with watching numbers slowly gather into a predictable pattern.  So I told the students it would be pretty cool if we could get more numbers.  Which is why I am going to start a page on here called the Coin Flip Experiment and whenever I have time I am going to do another 100 flips and add to the page to see when we in fact hit a percentage error of zero.

Imagine that, I just got goosebumps, a percentage error of zero.

I believe.

Source:  http://www.convergented.com/2013/04/heads-you-learn-tails-you-learn/