Randomization and Bias

Collecting Data A population is a group of people or objects that you want information about. Information can be gathered from a sample (subset) of the population. It may be too difficult, time-consuming, or expensive to survey everyone in the population.

Selecting a Sample Method SELF-SELECTED SAMPLE Members of a population volunteer to be in the sample. SYSTEMATIC SAMPLE A rule is used to select members of a population. Selecting every other person. CONVENIENCE SAMPLE Easy-to-reach members of a population are selected. Selecting those in the front row. RANDOM SAMPLE Each member of a population has an equal chance of being selected.

Why is randomization so important? Random sampling reduces bias by giving all individuals an equal chance to be chosen Why is randomization so important?

Different Randomized Designs Completely randomized design – all experimental units have the same probability of being selected for application of the treatment Randomized block design – involves first grouping experimental units according to a common characteristic, and then using random assignment within each group Matched pairs design – creating blocks that are pairs, in each pair, one unit is randomly assigned the treatment. Single-blind study – either person conducting the experiment or the subjects have knowledge of the treatment, but not both Double-blind study – neither the person conducting the experiment nor the subjects have knowledge of the treatment Different Randomized Designs

Self-Selected, Convenience, Systematic, or Random? A sportswriter wants to survey college baseball coaches about whether they think wooden bats should be mandatory throughout college baseball. Identify the type of sample described. The sportswriter contacts only the coaches that he has cell phone numbers for in order to get quick responses. The sportswriter mails out surveys to all the coaches and uses only the surveys that are returned. Self-Selected, Convenience, Systematic, or Random?

Convenience sample: the sportswriter selected coaches that were easily accessible to him. Self-selected sample: the coaches can choose whether or not to respond. Solution

“Do you think helmets should be required by law for bike riders “Do you think helmets should be required by law for bike riders?” “Do you think helmets should be required by law to save the lives of bike riders?” How are these questions worded differently? How is this related to bias? Besides how you pick your sample, what else can make your study biased?

A good sample is representative A good sample is representative. This means that each sample point represents the attributes of a known number of population elements. Bias often occurs when the survey sample does not accurately represent the population. The methods you use to collect data may also create bias, not just the sample of people/objects. Biased/Unbiased

Types of Bias Undercoverage Bias Systematically excluding specific parts of the population Causes over or under representation in the sample Non Response Bias Responses are not obtained from those selected for the sample Usually because they refuse to participate Response Bias Response bias can also be caused by how the administrator of the survey looks or acts People lie Wording of the Question Response bias can also be cause by question wording which may favor one outcome over another

Again, the way a question is worded could result in response bias Again, the way a question is worded could result in response bias. Here are some reasons why a question may be biased: Encourages respondent to answer in a particular way by using a leading question Perceived as too sensitive to answer truthfully Respondent is not provided with enough information to give an accurate opinion Creating Questions

Think of something you would like to know about your classmates Think of something you would like to know about your classmates. For example, favorite tv show or favorite color. To save time, you are not going to ask everyone, but instead ask a sample of students. Name a biased way to select a sample Name an unbiased way to select a sample Name a biased way to ask the question Name an unbiased way to ask the question Groups of 3-4