Students struggled with…

  • Explaining statistically significant with greater detail than “Yes because it is below 5%.”
  • Thinking that inference about cause-and-effect needs both random assignment and random samples.  Many students said an experiment that had random assignment but not a random sample could not determine causation because they couldn’t generalize it to the whole population.

Remedies…

  • Part of this issue will resolve itself throughout the year.  This is the first time students have to describe statistical significance so some error is not surprising.  Make sure to give written feedback for their responses by adding in parts that are missing.  This will help students know what you expect.
  • Explain that it is common in real experiments to not have a random sample.  The subjects are usually volunteers because if you did a random sample it would be very hard to convince people to come and be a part of your experiment.  So it is actually very common to have random assignment without random sample and that doesn’t mean we can’t determine causation.

Previous Day | Next Day