When I was getting my doctorate, I had a a professor who talked about the beauty of quantitative analysis (statistics) as beautiful. I love the research, too, but when I actually did my statistics coursework, I found it kind of dry and boring. We learned the mean, median mode of course as well as statistical analysis such as linear regression and chi squares.

Statistics and data analysis are now prominent in the common core standards, and so I thought about how the introduction to statistics could be more fun than in my first statistics class.  At the time, I was  building a ratio unit using Mondrian art, and I realized that the Mondrian squares students were designing were ideal for visualizing basic statistical analysis.  Students could literally see the mean, median and mode.

I wanted students to be able to understand why we want a larger sample size than 1 or 2.  For example, in my first statistics course I was told that the minimal acceptable sample size was 30. Got it. But why?! Well, in short with 30 you get a fairly consistent bell curve. But I hadn't really tested it.

I know that we learn better by experiencing, so I thought how can students prove that larger sample sizes are better for predicting outcomes for a population. In the project students collect and analyze different sample sizes to discover for themselves the reason for "more"data. And with Mondrian they can visually see it. We can say students in our class prefer yellow squares, and we can see it when we look at all the squares together. And for my class, statistics became a bit more tangible.