Teach 180: Why Randomization Matters (Day 10)

Today in Probability and Statistics we reviewed the article The Immediate Impact of Different Types of Television on Young Children's Executive Function from The American Academy of Pediatrics.  This article looked at four different tasks to measure executive function across the three treatment groups: fast-paced television, educational television and drawing.  Children were randomly assigned to one of the treatment groups and a table from the article (seen below) shows that the groups were fairly similar.  But does randomization always matter?  Will randomization tend to create groups that are similar?

To see if this was the case, we worked through an activity in the CollegeBoard's Curriculum Module on Random Sampling and Random Assignment.  There are 14 subjects listed on cards.  Each card has the subject's name, gender and IQ score.  Students shuffled the cards and split the cards into two piles of 7 cards.  Cards in the left pile represented the students that were randomly assigned to watch an episode of Sponge Bob Squarepants.  Cards in the right pile represented the students that were randomly assigned to the drawing group.

Next, students calculated the difference in the proportion of females in the two piles (Sponge Bob group - Drawing group) and the difference in the average IQ scores in the two piles (Sponge Bob group - Drawing group).  Each pair did this a total of five times and then we recorded our results on a dotplot.  Students noted that there was variability from sample to sample, but that the dots tended to center around 0 for both the difference in the proportions of females and the difference in the mean IQ.  Here you can see our class dotplot.

We finished the activity by watching a video done by Doug Tyson which uses Fathom to do the simulation.  If students weren't convinced by our dotplots that randomization matters, they were definitely convinced by the end of the video.