1
$\begingroup$

For my linear regression class we need to come up with some kind of simple experiment (we must collect the data, can't get it online) for which we can use the methods we have studied in class (simple linear regression, multiple linear regression, ANOVA, ...) to analyze the data.

I am having the hardest time coming up with something. Anyone have a good suggestion?

Last year someone did something silly like compare how fast mustard descends an incline compared to ketchup.

EDIT: Thanks for all the helpful ideas guys :) I ended up with doing some physical experiments on myself on a stationary bike (max speeds/power output at different gear ratios). Probably not the greatest experiment, but it utilizes something I enjoy.

$\endgroup$
1
$\begingroup$

Oh there are tons of things! I don't want to give you exact answers because ... well, I want to let you have some fun!

But... how do people's political views (e.g. rate someone on a 1 to 100 scale) relate to ... well whatever!

How do people's physical measurement relate to each other?

What do test grades depend on?

and so on!

$\endgroup$
1
$\begingroup$

Craigslist Missed Connections? Seriously, lots of interesting data there, fairly easy to get at, and it seems under-explored. Here's one neat example to inspire. It's not quite as rich as the OK Trends data, but it's available. One example might be how well the demographics mirror the census data for your city.

$\endgroup$
0
$\begingroup$

Inspired by @Dimitriy how about regressing daily-number of 'Craigslist Personals' versus temperature or local football team stats.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.