MATH 1040 - Introduction to Statistics

Lesson 3.1 Observational Studies

Reading

Reading sections are from the Introductory Statistics Textbook

Lesson

An Observational Study is a statistical study where the researcher simply gathers information.

Examples:

After the information is gathered, we simply make graphs and do calculations to analyze the data and to make conclusions.


No study is perfect. No matter how detailed we organize the variables and the data, there are outside influences. For example, if you ask for a favorite food, the answer could be influenced by allergies. If you ask analyze driving behavior, an individual’s driving could be influenced by where they grew up (city-driving vs. country-driving).

Confounding variables (also known as lurking variables) are other variables that can have an unintended influence on the data.

To avoid influences from confounding variables, we have to be careful about what variables we choose to gather. We can add variables that will help us gain insights to these extra variables.

In addition to eliminating confounding variables, the extra variables will give us a natural way to stratify or cluster the data.

Practice

Here are some examples of possible problems. Think of any possible confounding variables, then consider ways to get around these problems. (Note that your answers may not be the same as what is here. There are many possible confounding variables for any situation. The trick is to eliminate as many of them as possible.)

  1. Which stop sign is run through most often? You look at three different stop signs and count the number of cars that pass the stop sign, calculating the percentage of responses in support or against.
  2. Which candidate do you support? A party wants to determine the support they have, so they send surveys to ask people of their preferred candidate
  3. What is the most popular brand of car? You want to know which make of car is most popular in the area. So, you gather data from car dealerships in the area.