Statistics 10 Lecture 1 Introduction & Experiments

1. Most of us ask questions and make comparisons naturally and easily

Statistics can help us make rigorous and systematic comparisons of groups, their treatment, and responses and help us avoid arbitrary comparisons which cannot be generalized beyond the particular subjects being studied.

2. Good Comparisons require critical forethought (asking good questions)

Experiments and their design

• Treatment (an intervention)
• Control (what If we never intervened)
• A measurable response (a real outcome)

3. Even when we are clever, there can be problems

Confounding (vocabulary) -- the effect of an unforeseen characteristic, behavior, event or procedure on the response that cannot be distinguished from the proposed treatment.

4. But there are SOLUTIONS, things you can do to control problems

• Randomization(vocabulary) -- eliminates bias (vocabulary)
• Placebo (vocabulary) -- eliminates the "placebo effect" (vocabulary)
• Double Blind (vocabulary) -- eliminates bias
• Blind (vocabulary) -- may eliminate bias
• Replication (vocabulary) -- validate results

5. Chapter 1 goes over some actual applications

In class exercise, the navigation skills of men versus women. Let's begin with a famous quote:

Husbands think we should know where everything is—like the uterus is a tracking device. He asks me, "Roseanne, do we have any Cheetos left?" Like he can't go over to that sofa cushion and lift it himself.

---- TV SITCOM STAR ROSEANNE

6. Suggestions

Statistics involves a bit of vocabulary, you can always restate these in a way that helps you to understand the method of comparison and experiments.

Think about how you would conduct an experiment if you were in charge or if it suits you better: criticize an existing experiment. But do not be too harsh.