The Sampling Strategy section of your Research Strategy chapter (usually Chapter Three: Research Strategy) needs to be well structured. A good structure involves four steps: describing, explaining, stating and justifying. You need to: (1) describe what you are studying, including the units involved in your sample and the target population; (2) explain the types of sampling technique available to you; (3) state and describe the sampling strategy you used; and (4) justify your choice of sampling strategy. In this article, we explain each of these four steps:
First, the reader needs to know what you studied. This should include details about the following:
The units you measured (or examined).
Your target population.
If you used a probability sampling technique to select your sample, you will also need to describe:
Your sampling frame.
If you are unsure what of any of these terms mean (i.e., unit, sampling frame, population), you might want to read the article, Sampling: The basics, before reading on. If you feel comfortable with these terms, let's imagine we completed a dissertation on the career choices of students at the University of Oxford, England. Below we describe our units, target population and sampling frame (imagining that we used a probability sampling technique).
Career choices of students at the University of Oxford, England
We examined the career choices of all students at the University of Oxford, England. By all students we mean all undergraduate and postgraduate students, full-time and part-time, studying at the University of Oxford, England, enrolled as of 05 January 2011.
From this description, the reader learns the following:
Units: students
Population: all undergraduate and postgraduate students, full-time and part-time, at the University of Oxford, England
Sampling frame: all students enrolled at the University of Oxford, as of 05 January 2011 (i.e., according to Student Records, assuming this is the department that maintains a list of all students studying at the university)
Note the difference between the target population and the sampling frame, from which we select our sample (when using a probability sampling technique). They are the same in all respects apart from the fact that the sampling frame tells the reader that only those students enrolled in the university according to Student Records on a particular date (i.e., 05 January 2011) are being studied. If the list of students kept by Student Records is very different from the population of all students studying at the university, this should be made clear [see the article, Sampling: The basics, to understand more about sampling frames and potential sampling bias].
By the time you come to write up the Sampling Strategy section of your Research Strategy chapter, you should know whether the sampling frame is the same as the population. If it is not, you should highlight the difference between the two. This completes the first part of the Sampling Strategy section of your Research Strategy chapter.
Once you know what units you are studying, as well as your population and sampling frame, the reader will often want to know what types of sampling technique you could use. We say could use rather than should use because whilst there are certain ideal choices of sampling technique, there is seldom a right or wrong answer. Instead, researchers choose sampling techniques that they feel are most appropriate to their study, based on theoretical and practical reasons.
Broadly speaking, you could choose to select your sample from (a) your sampling frame using either a probability sampling technique (e.g., simple random sampling, systematic random sampling, stratified random sampling) or (b) from your population using a non-probability sampling technique (e.g., quota sampling, purposive sampling, convenience sampling, snowball sampling). To understand the differences between these techniques, as well as their advantages and disadvantages, you may want to start by reading the articles: Probability sampling and Non-probability sampling.
When explaining the types of sampling technique that were available to you in this part of your Sampling Strategy section, you should take into account: (a) the research strategy guiding your dissertation; and (b) theoretical and practical sampling issues.
The research strategy guiding your dissertation
Theoretically, the ideal sampling technique for a piece of research (i.e., probability or non-probability sampling) differs depending on whether you are using a quantitative, qualitative or mixed methods research design.
Theoretical and practical sampling issues
Whilst there are theoretical ideals when it comes to choosing a sampling technique to use for your dissertation (i.e., probability or non-probability sampling), it is often practical issues that determine not only whether you choose one type of sampling technique over another (e.g., non-probability sampling over probability sampling), but also the specific technique that you use (e.g., purposive sampling over quota sampling; i.e., both are non-probability sampling techniques). Such practical issues range from whether your target population is known (i.e., whether you can get access to a list of the population) to whether you have the time and money to get access to such a list [click on the relevant article to understand the advantages and disadvantages (i.e., theoretical and practical considerations) of the different probability sampling (e.g., simple random sampling, systematic random sampling, stratified random sampling) and non-probability sampling techniques (e.g., quota sampling, purposive sampling, self-selection sampling, convenience sampling, snowball sampling)].
Assuming that you understand the differences between these sampling techniques, and their relative merits, let's consider what sampling choices are open to us using our example of career choices of students at the University of Oxford, England. The green text illustrates what we have already written above.
Career choices of students at the University of Oxford, England
We examined the career choices of all students at the University of Oxford, England. By all students we mean all undergraduate and postgraduate students, full-time and part-time, studying at the University of Oxford, England, enrolled as of 05 January 2011.
Since our research drew on a quantitative research design, the ideal would have been to use a probability sampling technique because this allows us to make statistical inferences (i.e., generalisations) from our sample of students to all students at the university. Such a probability sampling technique would provide greater external validity for our findings. Since we wanted to compare the career choices of different strata (i.e., groups of students); more specifically, males and females, the appropriate choice of probability sampling technique would have been a stratified random sample. However, if it were not possible to use a probability sampling technique, we could have used a non-probability sampling technique. Since we wanted to compare different strata (i.e., groups of students) and achieve a sample that is as representative as possible of our population, we could have used a quota sample.
From this explanation, the reader learns the following:
Types of sampling strategy available: probability and non-probability sampling
Ideal choice: probability sampling
Preferred choice of probability sampling technique: stratified random sample
Preferred choice of non-probability sampling technique: quota sample
When you are writing up this part of the Sampling Strategy section of your Research Strategy chapter, you may be expected to include a much more comprehensive list of reasons why you prefer one type of sampling strategy (i.e., probability or non-probability) and more specifically, a particular sampling technique (e.g., stratified random sampling over quota sampling). We provide information about the advantages and disadvantages of these different sampling strategies and sampling techniques in the following articles: for probability sampling, see simple random sampling, systematic random sampling, stratified random sampling; for non-probability sampling techniques, see quota sampling, purposive sampling, self-selection sampling, convenience sampling, snowball sampling.
Third, you need to state what sampling strategy and sampling technique you used, describing what you did.
Again, let's consider this for our example of career choices of students at the University of Oxford, England. The green text illustrates what we have already written above.
Career choices of students at the University of Oxford, England
We examined the career choices of all students at the University of Oxford, England. By all students we mean all undergraduate and postgraduate students, full-time and part-time, studying at the University of Oxford, England, enrolled as of 05 January 2011.
Since our research drew on a quantitative research design, the ideal would have been to use a probability sampling technique because this allows us to make statistical inferences (i.e., generalisations) from our sample of students to all students at the university. Such a probability sampling technique would provide greater external validity for our findings. Since we wanted to compare the career choices of different strata (i.e., groups of students), including males and females, the appropriate choice of probability sampling technique would have been a stratified random sample. However, if it were not possible to use a probability sampling technique, we could have used a non-probability sampling technique. Since we wanted to compare different strata (i.e., groups of students) and achieve a sample that is as representative as possible of our population, we could have used a quota sample.
In the event, we used quota sampling to select the sample of students that would be invited to take part in our dissertation research. Student Records provided us with the appropriate quotas for male and female students, which showed a 53:47 male-female ration [NOTE: this is a fictitious figure]. We selected a sample size of 200 students, which was based on subjective judgement and practicalities of cost and time. Therefore, we sampled 106 male students (i.e., 53% of our sample size of 200 students) and 94 female students (i.e., 47% of our sample size of 200 students). For convenience, we stood outside the main library where we felt the thoroughfare (i.e., number of students passing by) would be highest.
From this statement and description, the reader learns the following:
Sampling strategy chosen: non-probability sampling
Specific sampling technique used: quota sampling
Details of quota sampling:
strata (i.e., groups of students) of interest are males and females
ratio of males-females at the university was 53:47
sample size selected was 200 students
quota sample filled based on ease of access to students at the main university library.
Again, when you are writing up this part of the Sampling Strategy section of your Research Strategy chapter, it may be appropriate to include greater description of the sampling technique you used.
Finally, you need to justify your choice of sampling strategy. When writing up the Sampling Strategy section of your Research Strategy chapter, you may find it easier to combine the third and fourth steps (i.e., stating and describing the sampling strategy you used, as well as justifying that choice). Taking our example of the career choices of students at the University of Oxford, England, we illustrate how the two steps can be integrated. As before, the green text illustrates what we have already written above.
Career choices of students at the University of Oxford, England
We examined the career choices of all students at the University of Oxford, England. By all students we mean all undergraduate and postgraduate students, full-time and part-time, studying at the University of Oxford, England, enrolled as of 05 January 2011.
Since our research drew on a quantitative research design, the ideal would have been to use a probability sampling technique because this allows us to make statistical inferences (i.e., generalisations) from our sample of students to all students at the university. Such a probability sampling technique would provide greater external validity for our findings. Since we wanted to compare the career choices of different strata (i.e., groups of students), including males and females, the appropriate choice of probability sampling technique would have been a stratified random sample. However, if it were not possible to use a probability sampling technique, we could have used a non-probability sampling technique. Since we wanted to compare different strata (i.e., groups of students) and achieve a sample that is as representative as possible of our population, we could have used a quota sample.
In the event, we used quota sampling to select the sample of students that would be invited to take part in our dissertation research. We were unable to use a stratified random sampling, our preferred choice, because we could not obtain permission from Student Records to access a complete list of all students at the university. Without any other way of attaining a list of all students, we had to use quota sampling. However, Student Records did provide us with the appropriate quotas for male and female students, which showed a 53:47 male-female ration [note: this is a fictitious figure]. We selected a sample size of 200 students, which was based on subjective judgement and practicalities of cost and time. Therefore, we sampled 106 male students (i.e., 53% of our sample size of 200 students) and 94 female students (i.e., 47% of our sample size of 200 students). For convenience, we stood outside the main library where we felt the thoroughfare (i.e., number of students passing by) would be highest.
From this justification, the reader learns the following:
Main reason for rejecting the ideal sampling strategy:
Access to a list of all students (i.e., the sampling frame needed for probability sampling) was not granted by Student Records.
No other way of attaining a list of all students was available.
When you think about justifying your choice of sampling technique when writing up the Sampling Strategy section of your Research Strategy chapter, you should consider both practical reasons (e.g., what time you have available, what access you have, etc.) and theoretical reasons (i.e., those relating to the specific sampling technique, but also your choice of research paradigm, research design and research methods).