d i g i t a l   a d v i s o r . . .

Choosing People to Participate in Your Study

The people who actually participate in your study will constitute a very tiny selection from the universe of people who could possibly participate. Obviously, they must be people who are willing to participate, and who will agree to be observed, interviewed, or surveyed or whatever.   How do you find them?  Should you ask everyone in the universe if they are willing to do that?   Probably not.  Instead, you will ask a set of people you have access to, either because you have a mailing list or because they are in your school or because they are students in your class.  You will ask a group of people that you can access.


The people who actually participate in your study are called your sample. The people you have access to, because they are in your school or because you have a mailing list, are your sampling frame.   They are the list who could have participated because you have the same access to them as you do to the people you actually selected. The people you are ultimately interested in are the population of interest. Rarely do we care only about the three teachers we observe or the 30 students we survey.  We care a larger population and we are using our sample to represent that larger group.  So we want our sample to reflect the population as a whole, and not to be idiosyncratic.  The method we use to select a sample, and the sampling frame we use, will influence how well the sample reflects the population as a whole.  Here are some of the ways people select samples.

1. Convenience Sampling

The most common, and least defensible, method of sampling is to sample by convenience

"I will study the kids in Ms Jones' class because I know her and I know she will allow me to hang around there and interfere with her routines and won't kick me out.

There is an implicit sampling frame here which is a list of all the teachers you know.  It is clearly an idiosyncratic frame.  And the particular sample, of students in Ms Jones' class, is a is probably not representative of all the teachers in your sampling frame.  You are picking them because of your relationship with Ms Jones.  What biases are introduced by using this sample?  When you finish the study, what kind of conclusions can you say about other students in other classrooms?

2.  Random Sampling

Random sampling is usually proposed as a solution to the problem of bias.  If you make a list of all the possible kids you could study, and select a sample at random, then all kids on the list have an equal chance of being selected, and your selection won't be biased if everyone has an equal chance of being selected. The list you make can still be focused on particular types of kids.  For instance, you could make a list of all ED kids who are in mainstream placements in the Lansing School District, and then choose 6 or 8 or 20 of them for your study. Then, at the close of your study, you can say for sure that your sample at least represents the list in your sampling frame, and you can speculate about how the sampling frame might be similar to or different from the larger population of ED kids.

3.  Purposely Sampling Different Situations or Types 

Purposive sampling is also intended to counteract the potential biases in convenience sampling.  For instance, if you want to study Ms Jones' class, you could also purposely select Ms Smith's class, knowing that you and she disagree on virtually every educational issue.  By purposely selecting such a second case, you force yourself to consider different types of circumstances and to test your ideas in different types of classroom settings.  These counter-cases are especially useful in case studies because they push your thinking and force you to be more explicit about how, or when, one thing leads to another.

4.  Sampling Alternative Points of View within a Single System

The idea behind this sampling idea is to insure that you get view points from people who live in  different parts of the situation.  It presumes the old adage that "where you stand depends on where you sit."  It also presumes that when you speak to people at the upper end of a hierarchy, you will get only their representation of what is going on, and it will be shaded to make them or their organization look better.  This applies to district superintendents, school principals, or classroom teachers.  So rather than taking the principal's word that the new discipline policy is working well, ask teachers about it, and ask students about it, and ask the custodian about it or the lunch room clerks.


© Mary Kennedy, 2006


Evaluating Research

Reasoning with Evidence

Doing Your Own Research