How do you select participants in a research study?

Most research within the social sciences, and that includes Deaf Studies, is conducted with people. These people who are involved in research studies are sometimes called subjects. Nowadays, this is considered to be a disempowering term, and has been replaced by participants. This is more than a politically correct alteration. You should view those upon (or with) whom you conduct research to be participating in the research enterprise. Without participants there is no research, and they often have an interest in the outcomes of the research. The only place where you can refer to them as subjects is in the Results section of your dissertation, because �subjects� is now only used as a statistical term.

Selecting who will participate in your study is a very important step in the research process, and requires careful thought. Indeed, it is something you need to think about early on in the process of designing your research study. It is easy to create an excellent research proposal, and then find that it is not possible to recruit the participants that the study requires. And starting over is never easy!

Let us use another example to illustrate the importance of selecting appropriate participants. Assume that you are interested in how Deaf people who are native sign language users process BSL information. You need to start by defining what you mean by a native user of BSL. This definition cannot be an arbitrary one � it is fundamentally related to The Research Question and The Hypothesis. If your interest is in brain processing, then your definition will have to bear in mind developmental aspects of language learning. You may decide to say that the Deaf people in your study must have Deaf parents and have learned to use BSL before the age of 6 years. This is called an operational definition. This means that for the purposes of your study this is what a Deaf native signer is. This allows readers to argue with your definition, and not be left in the dark about how you selected participants for involvement in your study. The definition is worded the way it is because we know that there are critical periods for language development. We also know that children require input from native users of a language if they are to learn it successfully and develop the necessary brain structures for processing the language. If we used a definition that all participants must be fluent adult users of BSL, then the study may include those who learned BSL much later in life and therefore could process the language in a very different way.

Once you settle upon a definition that determines who can be involved in your study, you have your target population. Obviously, you cannot select every Deaf person in the UK who has Deaf parents and learned BSL before age 6 years! So, you select a smaller group who you consider to be representative. Who are and are not representative depends upon your hypothesis and the nature of your study. For this reason it is best covered by your supervisor. Here we will consider how many participants you will require.

The number of participants depends upon the methodology you are using. If you have elected to use a case study approach, then only one person may be involved. For a focus group there may be only 5 or 6 people. Typically, qualitative methods require fewer participants. This is because you are collecting qualitative data, which is rich in meaning and complex in nature. For quantitative studies, especially where you wish to use statistical analysis, then you should aim for many more. It is not possible to give an exact guide as to the number of participants you should recruit, but the general maxim is �the more, the better�! As a general guide, for each group of participants you wish to compare you should try to recruit at least 20 participants. This requires a lot of planning and a lot of effort, so you need to make sure you leave yourself enough time to recruit enough participants and conduct the research.

Often, you may find that you are stuck, and cannot recruit as many participants as you would like. Are there ways around this? Yes, there are, but you simply cannot beat getting as many participants as possible! One way around the problem is to take more measurements, and more reliable measurements. This means that instead of asking only one question, ask twenty. And make sure that those questions are reliable and valid. Consider the following analogy. You have a tape measure, and have to find out the length of a certain make of desk. Now let us assume you have access to only one example of that brand of desk. If you can make only one measurement, then you want to be sure that your tape measure is reliable (would give the same length each time) and valid (measures using a scale you can trust). If you are allowed to make many measurements, then the reliability of the tape measure is not so important. You can take lots of measurements, and use the mean of those measurements, hopefully averaging out any errors due to using a dodgy tape measure. The best scenario is to make lots of measurements using a reliable tape measure! Clearly, none of these approaches is as good as having 30 or 40 desks that you can measure and average.

Once everything is in place, the planning side of the research is finished with. Now you have to conduct the research. Often this will involve conducting a pilot study. This will help you develop standardised instructions for participants, and highlight any problems in your design. These problems can then be eliminated before running the full study.

A pilot study is useful for many reasons, and should be conducted if at all possible. Pilot studies allow you to:

1.     Develop standardised instructions. See below.

1. Become familiar with the procedure. If you have been through the procedure of running your study a few times, it becomes easier to administer. This can make running the study quicker, and mean that the participant sees someone who looks like they know what they are doing!

3.     Check that the data are appropriate. After running a few pilot participants, you will also have some pilot data. Look at this data carefully. At this stage you can check whether analysing it will allow you to answer your hypothesis.

1. Identify procedural problems. Through piloting you may identify questions that participants do not understand or are prone to misinterpret. Computer software may crash if certain keys are pressed. Participants may get tired because the procedure is too long. Your instructions may be unclear and misunderstood. All such procedural problems can usually be overcome. Piloting allows you to identify them before it becomes too late.

If you are running a study, especially a quantitative piece of research, then it is vital that all participants receive the same instructions. How someone is told to do something, as well as what they are told to do, will affect how they respond. The wording (or signing) of instructions should be standardised and followed exactly for each participant. Sometimes a participant will ask you for clarification. You also need to be clear about the boundaries of the clarification � what are you prepared to tell the participant in answer to questions they have about the procedure?

Standardisation is also important in how you collect your data. If you interview 30 participants in BSL, and collect data from 10 more using e-mail, then can you combine the two sets of data? The way you have solicited information may affect the information you collect. So, if you do this, you must (a) report it in the Procedure section of the Method, and (b) check the data before you combine it.

Whatever procedure you use, you must treat your participants in an ethical manner. All proposals will be checked to ensure that they follow ethical guidelines. The guidelines used for this Unit will be those laid down by the British Psychological Association, a copy of which is appended to these course notes. Ethics in research is very important, so is considered in more depth in the next chapter of these notes.

You should take great care when recording your data. This is a potential source of error. It is no good making valid and reliable measurements, if you then make a mistake in recording them. Record data as soon as it is collected, and do not rely upon notes scribbled on a piece of rough paper. Although this may be expedient at the time, you will come to regret it later. Keep a folder in which you can keep questionnaires, transcriptions of video materials, or any other form of data you collect.

The data is the life-blood of your dissertation, and should be protected at all costs. It is advisable to make back-up copies of paper-based recordings. Better still, computerise your data and keep a back-up copy somewhere safe. This could take the form of a word processing file with transcriptions or tables with interview data, or even a spreadsheet with scores recorded in a transparent manner. Remember, you will often come back to your data after a break, and it is very frustrating to remember a coding system or what a column indicates of you have not written it down somewhere. Be meticulous in keeping records, and life will be a lot easier.

Once you have collected all of your data, it is time to analyse it. The exact kind of analysis you perform will be determined by the kind of data you collected and your hypothesis. At this stage you must consult with your supervisor about the most appropriate form of analysis to pursue.

What is the best method of selecting participants?

How are participants selected for quantitative research?

How are respondents selected in research?

What sampling techniques will be used to select the participants?