Making Learning Fun!

[vhawk01]

So, I have to give a small, 5 minute or so, presentation to a small group of my peers and an epidemiologist. He is...eccentric...and so has requested that, rather than lecturing on our selected topic, we come up with some sort of fun, engaging presentation. The class is Evidenced-based Medicine, which is sort of an intro to biostatistics and study design. My topic is Correlation.

I don't need any help explaining correlation to them. My classmates are mostly stats-challenged, and this is a fairly basic course. However, I'm not very creative, in general, and I have a hard time with projects like this. I'm wondering if anyone has some suggestions. I figured this is at least tangential to the forum, and I know some of you are professors and teachers, and most of you are or were students. These are all grad students, and most of them have little stats knowledge. Can anyone think of any fun way of presenting the concept of Correlation? We're allowed to do basically anything, with no specific requirements. We were instructed to make it a "Explaining it to your parents who don't know anything" level of complexity.

Appreciate any help you guys have.

[Duke]

Boil down what you want to say about correlation to one major theme, show a picture to demonstrate it, and invent some sort of activity that will make it clear.

For instance, if you want to stress that correlation does not imply a causal relationship, well, maybe track sizes of people, and also whether or not they pee standing up. They should come up with bigger people peeing standing up more of the time, so there would be some correlation. Then you point out that the actual cause is the sex of the individual, and not their size. It's not perfect, but it should get the idea across.

5 minutes is basically "no time at all" to teach something.

[vhawk01]

[ QUOTE ]
Boil down what you want to say about correlation to one major theme, show a picture to demonstrate it, and invent some sort of activity that will make it clear.

For instance, if you want to stress that correlation does not imply a causal relationship, well, maybe track sizes of people, and also whether or not they pee standing up. They should come up with bigger people peeing standing up more of the time, so there would be some correlation. Then you point out that the actual cause is the sex of the individual, and not their size. It's not perfect, but it should get the idea across.

5 minutes is basically "no time at all" to teach something.

[/ QUOTE ]

Right, I'm not really teaching it so much as bringing it up and highlighting it. There are 12 of us, each with our own topic, and this is all stuff that we are supposed to be familiar with. These are peers.

Thanks for the ideas.

[SoReady]

Bring it up and highlight it, hmmm.

"Correlations are essentially relationships.

It's been shown that CEO's, for whatever reason, tend to be tall. The average CEO in the USA is over 6 feet tall, IIRC.
In fact, there's a significant relationship between height and salary in the USA today. Tall people tend to make more money than shorter people.

There's a positive correlation between income and height.

Having picked up on this tendancy, this weird occurance, we can now guess - "why do tall people make more money?"
You know, you need to observe something before you can guess how it works. Correlations are a simple observation.

In this example, I think it was shown that humans naturally are attracted to, on some degree, tall people. Tall people garner a certain natural respect from others. That helps the giants progress faster in the business world than your average shorty, hence the higher salary."

I'd do something succinct and basic like that. *shrug*

Does it need to be an any-more-detailed definition than that?

[dknightx]

[ QUOTE ]
Bring it up and highlight it, hmmm.

"Correlations are essentially relationships.

It's been shown that CEO's, for whatever reason, tend to be tall. The average CEO in the USA is over 6 feet tall, IIRC.
In fact, there's a significant relationship between height and salary in the USA today. Tall people tend to make more money than shorter people.

There's a positive correlation between income and height.

Having picked up on this tendancy, this weird occurance, we can now guess - "why do tall people make more money?"
You know, you need to observe something before you can guess how it works. Correlations are a simple observation.

In this example, I think it was shown that humans naturally are attracted to, on some degree, tall people. Tall people garner a certain natural respect from others. That helps the giants progress faster in the business world than your average shorty, hence the higher salary."

I'd do something succinct and basic like that. *shrug*

Does it need to be an any-more-detailed definition than that?

[/ QUOTE ]

zzzzzzz

[dknightx]

what you should try to do is poke fun at how people can often times confuse correlation with causation. i'm sure you can think of some funny examples, and possible demostrate it yourself (maybe go around campus asking random people questions, and tape it).

[vhawk01]

[ QUOTE ]
Bring it up and highlight it, hmmm.

"Correlations are essentially relationships.

It's been shown that CEO's, for whatever reason, tend to be tall. The average CEO in the USA is over 6 feet tall, IIRC.
In fact, there's a significant relationship between height and salary in the USA today. Tall people tend to make more money than shorter people.

There's a positive correlation between income and height.

Having picked up on this tendancy, this weird occurance, we can now guess - "why do tall people make more money?"
You know, you need to observe something before you can guess how it works. Correlations are a simple observation.

In this example, I think it was shown that humans naturally are attracted to, on some degree, tall people. Tall people garner a certain natural respect from others. That helps the giants progress faster in the business world than your average shorty, hence the higher salary."

I'd do something succinct and basic like that. *shrug*

Does it need to be an any-more-detailed definition than that?

[/ QUOTE ]

Appreciate all the posts, everyone.

Not sure if this is a troll account or not, though. Awesome either way.

[vhawk01]

[ QUOTE ]
what you should try to do is poke fun at how people can often times confuse correlation with causation. i'm sure you can think of some funny examples, and possible demostrate it yourself (maybe go around campus asking random people questions, and tape it).

[/ QUOTE ]

Its funny, because my first instinct was to poke fun at people who think correlation DOESN'T imply causation. Its a mantra that is repeated by anyone with even a rudimentary knowledge of stats, especially when any new study comes out, but I don't think people really know what they are saying. Correlation doesn't imply causation, if by imply you mean prove. But it certainly does lend support to the idea of causation, and it makes that specific cause more likely to be the actual cause. It is always possible that there is some third thing that causes both, thus the correlation, but such is science...there will ALWAYS be possible additional causes. So, if two things are highly correlated, obviously you need to consider the direction of the causation, and consider any possible other factors which are driving the correlation, but it still really is evidence for causation.

[Neuge]

[ QUOTE ]
[ QUOTE ]
what you should try to do is poke fun at how people can often times confuse correlation with causation. i'm sure you can think of some funny examples, and possible demostrate it yourself (maybe go around campus asking random people questions, and tape it).

[/ QUOTE ]

Its funny, because my first instinct was to poke fun at people who think correlation DOESN'T imply causation. Its a mantra that is repeated by anyone with even a rudimentary knowledge of stats, especially when any new study comes out, but I don't think people really know what they are saying. Correlation doesn't imply causation, if by imply you mean prove. But it certainly does lend support to the idea of causation, and it makes that specific cause more likely to be the actual cause. It is always possible that there is some third thing that causes both, thus the correlation, but such is science...there will ALWAYS be possible additional causes. So, if two things are highly correlated, obviously you need to consider the direction of the causation, and consider any possible other factors which are driving the correlation, but it still really is evidence for causation.

[/ QUOTE ]

The two examples given in this thread are a perfect dichotomy to explain this. In the first example height and peeing standing up are correlated, but being tall obviously isn't the cause. Statistically being male causes one to be taller and is obviously the cause for the standers. In the second, height and salary are correlated (or at least perceived salary), and height is the cause. While there may be sociological and evolutionary explanations as to why it's the cause, those explanations don't cause the above average height itself.

[vhawk01]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
what you should try to do is poke fun at how people can often times confuse correlation with causation. i'm sure you can think of some funny examples, and possible demostrate it yourself (maybe go around campus asking random people questions, and tape it).

[/ QUOTE ]

Its funny, because my first instinct was to poke fun at people who think correlation DOESN'T imply causation. Its a mantra that is repeated by anyone with even a rudimentary knowledge of stats, especially when any new study comes out, but I don't think people really know what they are saying. Correlation doesn't imply causation, if by imply you mean prove. But it certainly does lend support to the idea of causation, and it makes that specific cause more likely to be the actual cause. It is always possible that there is some third thing that causes both, thus the correlation, but such is science...there will ALWAYS be possible additional causes. So, if two things are highly correlated, obviously you need to consider the direction of the causation, and consider any possible other factors which are driving the correlation, but it still really is evidence for causation.

[/ QUOTE ]

The two examples given in this thread are a perfect dichotomy to explain this. In the first example height and peeing standing up are correlated, but being tall obviously isn't the cause. Statistically being male causes one to be taller and is obviously the cause for the standers. In the second, height and salary are correlated (or at least perceived salary), and height is the cause. While there may be sociological and evolutionary explanations as to why it's the cause, those explanations don't cause the above average height itself.

[/ QUOTE ]

Good point. My biggest problem is that, when I said my prof is eccentric, I meant like Nathan Lane eccentric. He's really over the top, and put a lot of pressure on us to make this fun and creative and interactive. Fun, creative and interactive are about the three most dreaded words when it comes to me and projects/presentations. I haven't decided which is better:

Do a normal, straight presentation and then hit him in the face with a water balloon and make sound effect noises

or

Hit him with the waterballoon first.