It doesn’t take long after turning on the TV or going to a website before you see statistics and survey results. Think about surveys that you may see online…which ones do you answer? The ones you choose to answer are most likely ones that you have some interest in and feel strongly towards. If that’s the case, we need to be careful about how much weight to put in those survey results since the majority of respondents are ones who tend to go to one extreme or the other. Consider a survey done inside a Target on a Wednesday morning. There are certain groups of people who are more likely to be at a Target on a Wednesday morning so this will skew your results.
There is a reason that research companies are paid a lot of money to conduct unbiased studies. It takes significant work and thoughtfulness to obtain results that are free from bias. Before we look at some of the numerical strategies in Statistics, we will first look at some methods of sampling and potential biases that can come into play.
The population refers to the entire group being studied. This can be a very large group such as all Americans or it can be much smaller like juniors who are enrolled in an AP Statistics class at your school. A sample is a part/subset of the population. This sample is the group that will actually be studied and the goal is to make it as representative of the population as possible. Bias happens when the sample is not representative of the population and/or the method of sampling affects the results. In simplest terms, the method of sampling is to learn about the whole (or population) by studying a part (or sample).
There are several methods of sampling that can happen. This list is not exhaustive but does include the most common.
Simple Random Sampling is when everyone in the population has an equal chance of being studied. Examples of this could be a random number generator, random name generator, or picking names from a hat.
Systematic Random Sampling is a sample that follow a rule (something like every nth person). Examples of this would include selecting every 10th person on a list or every 7th person who walks by.
Stratified Random Sampling happens when the population is first divided into two or more groups based on certain characteristics. Then a simple random sample is taken from every group. The characteristics could be gender, age, race, political affiliation, etc.
Cluster Random Sampling starts with the population being divided into geographic groups. One or more of those groups are randomly selected and everyone in this selected groups are studied. The geographic groups could be cities, neighborhoods, classrooms, floors in a building, etc.
Below are two other sampling methods that are not random.
Convenience Sampling is a sample taken of people that are accessible to the surveyor. Only easy-to-reach members of the population are selected. An example of this would be if your population is Naperville residents, and your sample are residents in your neighborhood. Your neighbors are easy for you to survey but they are not representative of the entire population.
Self-Selected/Volunteer Sampling is a sample of volunteers from the population. At the beginning we talked about surveys that are done online where you can choose to respond. That would be an example of self-selected sampling.
There is a reason that research companies are paid a lot of money to conduct unbiased studies. It takes significant work and thoughtfulness to obtain results that are free from bias. Before we look at some of the numerical strategies in Statistics, we will first look at some methods of sampling and potential biases that can come into play.
The population refers to the entire group being studied. This can be a very large group such as all Americans or it can be much smaller like juniors who are enrolled in an AP Statistics class at your school. A sample is a part/subset of the population. This sample is the group that will actually be studied and the goal is to make it as representative of the population as possible. Bias happens when the sample is not representative of the population and/or the method of sampling affects the results. In simplest terms, the method of sampling is to learn about the whole (or population) by studying a part (or sample).
There are several methods of sampling that can happen. This list is not exhaustive but does include the most common.
Simple Random Sampling is when everyone in the population has an equal chance of being studied. Examples of this could be a random number generator, random name generator, or picking names from a hat.
Systematic Random Sampling is a sample that follow a rule (something like every nth person). Examples of this would include selecting every 10th person on a list or every 7th person who walks by.
Stratified Random Sampling happens when the population is first divided into two or more groups based on certain characteristics. Then a simple random sample is taken from every group. The characteristics could be gender, age, race, political affiliation, etc.
Cluster Random Sampling starts with the population being divided into geographic groups. One or more of those groups are randomly selected and everyone in this selected groups are studied. The geographic groups could be cities, neighborhoods, classrooms, floors in a building, etc.
Below are two other sampling methods that are not random.
Convenience Sampling is a sample taken of people that are accessible to the surveyor. Only easy-to-reach members of the population are selected. An example of this would be if your population is Naperville residents, and your sample are residents in your neighborhood. Your neighbors are easy for you to survey but they are not representative of the entire population.
Self-Selected/Volunteer Sampling is a sample of volunteers from the population. At the beginning we talked about surveys that are done online where you can choose to respond. That would be an example of self-selected sampling.
Quick Check
You want to conduct your own survey to find out how many hours per week students at your school spend in extra-curricular activities. Choose the sampling method that is described in each of the following.
a) You obtain a list of all of the students, and you survey every 20th student on that list.
b) You survey all of the students in each of your classes throughout the day.
c) Five random 3rd hour classrooms are selected and all students in those classes are surveyed.
d) 100 students are randomly selected from each of the class years.
e) You input each student’s ID number into a random number generator and survey the first 50 students that the generator selects.
f) You set up a table in the cafeteria with a sign that says Extra-Curricular Survey, and students come to the table if they choose.
Quick Check Solutions
You want to conduct your own survey to find out how many hours per week students at your school spend in extra-curricular activities. Choose the sampling method that is described in each of the following.
a) You obtain a list of all of the students, and you survey every 20th student on that list.
b) You survey all of the students in each of your classes throughout the day.
c) Five random 3rd hour classrooms are selected and all students in those classes are surveyed.
d) 100 students are randomly selected from each of the class years.
e) You input each student’s ID number into a random number generator and survey the first 50 students that the generator selects.
f) You set up a table in the cafeteria with a sign that says Extra-Curricular Survey, and students come to the table if they choose.
Quick Check Solutions
Each of these sampling methods, no matter how well done, has the potential to lead to bias. Some of these reasons are listed below, but remember that just like with the sampling methods, this list does not cover every type of bias that can occur.
When does bias occur in sampling?
How can bias play a role in these sampling methods?
When does bias occur in sampling?
- Volunteer-Response Surveys: typically only those passionate enough to respond do.
- Undercoverage: the sample is not representative of the population.
- Non-Response: does not take into account the opinions of those who choose not to respond in self-selected sampling.
- Wording of the Questions: can influence how people respond.
How can bias play a role in these sampling methods?
- Simple Random Sampling or Systematic Random Sampling: you may not get a representative sample.
- Stratified Random Sampling: groups may not be equal in influence.
- Convenience Sampling: typically leads to undercoverage.
- Self-Selected or Volunteer Sampling: may only get results from those passionate about the topic.