STEP TWO
Choose your sampling strategy and sampling technique

Once you understand the basic principles and characteristics of sampling, you need to start thinking about the overall sampling strategy that you will use to collect the data needed for your dissertation. In other words, you need to identify: (a) the units you will be studying (i.e., whether they people, objects, cases, pieces of data, etc.), and their characteristics; (b) the sampling technique that you will use to select the units that you will include in your sample (i.e., whether a probability or non-probability sampling technique, and the specific technique being used; e.g., a simple random sample as your choice of probability sampling technique); and (c) the practical aspects of selecting the units using this sampling technique (e.g., creating rules for randomly sampling participants within an overt, structured observation).

When you choose your sampling strategy, you also need to take into account: (a) the consistency between your sampling strategy, research design and research methods; (b) your chosen route, and the approach you have adopted within that route; and related to this; and finally, (c) the practicalities of choosing such a sampling strategy for your dissertation (e.g., what time you have available, what access you have, etc.).

The consistency of your sampling strategy with your research strategy

One of the major aims of quantitative research (i.e., quantitative dissertations) is the desire to make generalisations (i.e., statistical inferences) from the sample you are studying to the broader population you are interested in (e.g., from the 300 students in your sample to the 15,000 students at your university). To make such generalisations, you need to make sure that the units you are studying in your sample (e.g., statistical inferences, the 300 students) are representative of this broader population (i.e., representative in terms of the characteristics of the units, such as age, gender, educational background, etc.). The more representative your sample is of the broader population, the greater the external validity (i.e., generalizability) of your findings (NOTE: We discuss external validity further in STEP FIVE: Research quality). To achieve such representativeness, the units that end up in your sample should be drawn at random from the population, which is a major aspect of probability sampling techniques. However, this is not always possible, such that non-probability sampling techniques have to be used instead, reducing the potential representativeness of your sample, and limiting the external validity of your findings.

Added to this, you have to consider the consistency of your sampling strategy in terms of your research strategy. For example, if you are following an experimental research design, participants need to be randomly assigned to groups using a probability sampling technique (e.g., a simple random sample). However, if this is not possible, you will only be able to follow a quasi-experimental research design, which lacks some of the explanatory power of experimental research designs. Therefore, when considering the sampling strategy you want to follow, you need to make sure that it is consistent with your research strategy. If, as the above example illustrates, this is not possible, you may not only have to change your desired sampling technique (e.g., from a probability sampling one to a non-probability sampling one), but also go back and change an aspect of your research strategy (e.g., a change of your research design from an experimental one to a quasi-experimental one). When such changes have to be made, go back to the relevant section within STAGE SIX: Research strategy, and work back through the steps. For example, if you have to change your research design from an experimental to a quasi-experimental one, you would need to read up on quasi-experimental research designs and how they are different from experimental research designs. This is important because as you work your way through STAGE SIX: Research strategy, you find that many things are influenced by such changes, especially factors relating to research quality, which we discuss in STEP FIVE: Research quality.

Your chosen route and approach

Since you are taking on a Route #1: Replication-based dissertation, the main journal article you are interested in may be able to give you some idea of the sampling strategy you could follow in your dissertation. However, the extent to which this is the case will vary depending on the route you adopt, and the approach within that route. Route A: Duplication, Route B: Generalisation and Route C: Extension are discussed in turn:

• Route A: Duplication

  If you are pursuing an approach to Route A: Duplication where you are using the original data collected by the authors in the main journal article, there is no need to set a sampling strategy in the traditional sense (i.e., you are not collecting data). However, you will need to focus on possible weaknesses in the sampling strategy adopted in the main journal article in order to assess possible threats to internal validity and external validity. To do this, you need to first assess the sampling strategy that was used in the main journal article in terms of the sampling technique that was applied (i.e., was it a probability or non-probability sample, and what type of probability or non-probability sampling technique was used?). Second, assess whether this sampling technique is consistent with the research design used in the main journal article (e.g., a non-probability sampling technique such as self-selection sampling would, in principle, be inconsistent with an experimental research design, since such a research design requires participants to be randomly assigned to groups, often based on specific sampling criteria). Third, if the sampling strategy used is inconsistent with the research design that was followed, read up on the advantages and disadvantages of the sampling technique that was used in the main journal article, as well as the sampling technique that would have ideally be used. You can do this in the Sampling Strategy section of the Fundamentals part of Lærd Dissertation. If the sampling strategy was consistent with the research design that was applied, you simply need to consider the advantages and disadvantages of the sampling technique that was used in the main journal article. When considering such advantages and disadvantages, you will also need to reflect on possible threats to internal validity and external validity resulting from the choice of sampling strategy. We discuss this further in STEP FIVE: Research quality.

  If you are pursuing Route A: Duplication, and you are collecting your own data, rather than using the original data collected by the authors in the main journal article, you should be able to apply the sampling strategy used in the main journal article in your dissertation. However, there are some exceptions to this, including (a) an inability to apply the same sampling strategy for practical reasons, and (b) an ability to use a superior sampling technique to the one in the main journal article. In the case of (a), the authors may have been able to get access to a list of the population, which allowed them to use a probability sampling technique, whereas you were not. Alternately, the authors may have had the resources (i.e., time and money) to build a large sample (e.g., 500 units), which allowed them to stratify the sample into multiple groups based on a range of sampling characteristics (e.g., age, gender, educational background), but you are simply unable to attain such a large sample in order to achieve representative stratifications of the population (i.e., you cannot get an accurate proportion or number of a given gender, age, or educational background, etc.). In the case of (b), the authors in the main journal article may have been unable to use a probability sampling technique to create their sample, perhaps because they could not get access to a list of the population being studied, whereas you are able to do this. Therefore, you have an opportunity to create a more representative sample of the population, which will increase the potential external validity (i.e., generalizability) of your findings.

• Route B: Generalisation and Route C: Extension

  In dissertations based on Route B: Generalisation and Route C: Extension, you have to rely less on the sampling strategy used in the main journal article and focus more on the population and setting/context that you are interested in. After all, the characteristics of the new population or setting/context that are important are likely to be different from those characteristics that were important in the main journal article. For example, think about the important characteristics of a sample required to examine service quality amongst banking customers in the UK compared to hospital patients in the UK. In both cases, you may want to stratify your sample according to criteria such as age and gender, but in the study involving banking customers, you may also want to stratify your sample based on income level, whereas this would be less relevant amongst hospital patients because healthcare in the UK is free. Therefore, when creating your sample in a dissertation based on Route B: Generalisation or Route C: Extension, think about what the most important characteristics of that sample are in order for it to be representative of the population you are interested in.

The practicalities of your chosen sampling strategy

At the end of the day, the sampling strategy that you select is something very practical. For example:

• If you follow a probability sampling approach, you need to gain access to a complete list of the population that you are interested in from which you draw your sample. This means you have to go out and get such a list, if one even exists. If no single list does exist, you may have to bring together numerous sub-lists to create a final list from which you want to select your sample. You have to assign numbers to the list, find random numbers, before selecting your sample based on your sampling criteria (e.g., if you were using a stratified random sample - a probability sampling techniques - you would have to stratify your list according to the sampling characteristics you were looking for; perhaps a certain male-female ratio, educational level, and so forth).

• If you are following a non-probability sampling, such self-selection sampling, you need to let potential applicants or organisations know about your study. This will involve some kind of advertising or promotion, whether print media, the radio, an online notice board, or some other medium. The invitation will also need to follow certain ethical guidelines, making it clear what the study involves, but also more practical information, such as the types of applicant that are required (e.g., age, gender, or some other more subject-specific criteria).

All of these types of practicalities need to be taken into account when choosing your sampling strategy.

STEP THREE
Consider the practical aspects of your dissertation, and the implications that these may have for your sampling strategy

There are many ideals when it comes to setting a research strategy for your dissertation, and the sampling strategy is no different. As you will have noticed in the previous sections, such ideals may be based on (a) the sampling strategy adopted in the main journal article, (b) your chosen route, and the approach within that route, and (c) the sampling technique that is consistent with your chosen research design and research method.

However, it is often not possible to adhere to such ideals, especially at the undergraduate and master's level, for a variety of reasons. For example, probability sampling techniques require that you can attain a list of the population you are studying from which you can draw your sample. However, even if a list is readily available, it may be challenging to gain access to that list. The list may be protected by privacy policies or require a lengthy process to attain permissions. There may be no single list detailing the population you are interested in. As a result, it may be difficult and time consuming to bring together numerous sub-lists to create a final list from which you want to select your sample. As an undergraduate and master's level dissertation student, you may simply not have sufficient time to do this. At the same time, many lists will not be in the public domain and their purchase may be expensive; at least in terms of the research funds of a typical undergraduate or master's level dissertation student.

These and other issues highlight just some of the problems that you can face when devising the ideal sampling strategy for your dissertation. As a result, you often have to be pragmatic when setting your sampling strategy, devising the best sampling strategy that you can based on your available resources, whilst acknowledging in your write up (i.e., the Sampling Strategy section of your Chapter Three: Research Strategy) the weaknesses of your chosen approach.