Snapshot: A brief overview of considerations and techniques to use for gathering research data, especially in social sciences contexts
Gathering data should be one of the most exciting parts of any research project. But gathering the right kind of data, ready to be analysed in a way which will be useful and enlightening for your project can be a real challenge. Thinking about what you need to achieve this is really important before you start doing any data gathering at all. The key thing to think about right at the start is what kinds of questions you want your research to answer, in relation to any hypothesis you may have. This is the beginning of your research design.
Really, data is a fun concept. All that information you can play about with: categorise, convert into numbers, work out probabilities, standard deviations, means, medians, averages. But maybe it sounds a bit daunting, for example, do you need to know a lot of maths? What if you just want to interview some people? Start at the beginning and ask yourself some questions:
What people say, do, behave like or respond like are all possibilities for producing data in your project if it's a social science based piece of work (in fact, even in science based projects, what people do or say is of great interest). We'll look at a few basic variations on how data can be gathered from people, and look at each one in a bit more detail, as well as considerations on WHO is taking part.
Demographics - Basic considerations for all research involving people are many or all of the following: age, gender, economic social class, nationality, ethnicity, cultural background, language, location, employment status, health, literacy level, digital literacy level.
Ethics - Very important considerations surrounding consent, anonymity and privacy are involved in ALL research projects. Use of 'consent forms', acknowledgement of assured anonymity between respondent and data they produce, and privacy of all respondents taking part unless otherwise specifically stated or requested with their individual consent.
Sampling and Grouping - You need to consider how you will sample your participants, and how you will group them for your study. See below in the Research Design tab for more information on this.
Interviews - Usually one-to-one conversations between the interviewer and the interviewee. These can either be non-scripted, semi-scripted, or strict sets of questions.
Focus Groups - A group of people who are invited to a meeting to hold a conversation around some topic area, often with some specific ground to be covered. Most often the questions will be 'semi-scripted' rather than tightly controlled, to allow the participants to cover ground they feel is relevant to the topic being discussed.
Questionnaires - These are most likely done either over the phone or online. It's rare to use a strict questionnaire in a face-to-face setting.
Observation - Watching your participants, either in controlled surroundings, or 'in-situ', from a more distant perspective. Sometimes observation can include the researcher themselves being part of the observed group.
Probe don't Prompt - Never put words in your respondent's mouths. It's the easiest thing to say ".. you know what i mean..." or "...you mean such-and-such...". Refrain from doing this. Instead of using suggestions, ask "...can you clarify that a bit?" or "...could you give an example of what you mean?"
Recording and Precising - There's really only one way of how to remember what's been said, and that's to record the interview or discussion. In your research document, you'll have to write it up too, as your reader cannot be expected to listen to or watch the proceedings. That's known as precising, and means writing a fairly accurate (though not completely verbose) account of what happened.
Managing a group of people in a focus group - It can be difficult to allow everyone to speak, to control the most talkative in order to allow the less talkative to get a word in, to get people to speak clearly, slow enough for the recorder, and not all speak at once. It can be difficult to get people to say anything at all. Your experience and techniques as an interviewer will develop with practise, but don't be shy, and don't be too formal.
Keeping the conversation focused - Keeping on topic is always a challenge. Keep your crib sheet nearby, and refer to it to make sure you are covering the ground you want to find out about, but bearing in mind that you do want the conversation to flow naturally and perhaps bring out aspects you may not have thought of.
How many Questions? The rule we often follow in market and newspaper research is keep it short and this applies to all questionnaires. If you were doing the survey, you wouldn't want it to go on forever. If you need a lot of information, break it up into several shorter surveys, and make it easy for people to take part - online, or with easy to fill in professionally presented paper questionnaires.
What sort of Questions? It's always a challenge to think of the right questions to get your data, but the real skill comes in what sort of answers you offer. Refer to the 'Helpful Links' tab for wikipedia pages which give a lot of helpful information on this.
Open versus Closed? Open ended questions sound great in theory, but in practise often mean that noone tells you anything. people just don't fill in open ended questions unless they really have to, and they never give sufficient detail. Consider carefully when you will use a truly open ended question.
Advantages and disadvantages of multichoice and Likert Scales - Using predetermined sets of answer choices is often a much more productive way of gathering quantitative data, but isn't always suitable for in depth complex research, so again, careful planning in relation to what you're trying to find out is really worth it at the start of the project. Remember, you can use a mixture of question types, as well as types of data to build your research data picture.
In a controlled environment or an open environment - If you observe people, for example their behaviour, or doing things you want to research, you may wish to observe them in a controlled environment, or in their 'natural habitat'. Think about what's best for what you want to find out, and if you want to compare with any other type of environment.
In real life or online - Observing people's actions and behaviour whilst online is a growing aspect of research, for example to help with human computer interaction, application or game design, or digital literacy issues, and may require software or specialist techniques. Real life observation has its own challenges, for example controlling outside 'interference' or influences which may skew results of the research.
Sampling and control groups - Sometimes getting people to take part in observation research is more difficult, as time and location constraints can limit participant availability. Also, ensuring privacy in an observed setting can be especially problematic and may put some potential participants off.
Download the PDFs for some deeper reading on some of what we've discussed, and use the tab sections below to get a bit more understanding on terms, theory and techniques for your research.
Questionnaires and surveys
http://en.wikipedia.org/wiki/Questionnaire_construction Constructing a questionnaire
http://en.wikipedia.org/wiki/Quantitative_marketing_research Information on quantitative market research techniques
http://www.questionpro.com/a/showArticle.do?articleID=survey-questions Types of questions clarified
http://www.statcan.gc.ca/edu/power-pouvoir/ch2/exer/5214909-eng.htm#link02 An exercise to help develop a good questionnaire
Social Media for gathering data
Methodology and Methods
Research Design is about how the data is gathered in relation to how the data will be analysed. It is the setting, the context, the grouping of the participants, ethical considerations and the type of data which will be produced. (It's sometimes called experimental design for science based research.)
Things typically affected by research design choices are:
How are you going to get your data into manageable chunks to meaningfully analyse? Matching up your question to suitable mthods is the crux of what we're thinking about.
Qualitative data is any data you end up with that is long-form - talking, looking, patterns of behaviour over time, comparison systems, interpretive categorisation etc. It's often associated with data which isn't numbers. It's about depth and quality.
Quantitative data is usually numbers. It can be larger sets of numbers, but doesn't have to be. The quantity is in the numbers you're grouping together, not necessarily the amount of numbers. So 'of our sample of 50, 29 people said they like cheese' is a qunatitaive data statement. You can 'quantify' qualitative data, by assigning numbers to results, for example counting the amount of people who mentioned something in an opene ended response without being prompted to do so.
Mixing it Up
Mixed methods can be really useful to compare and contrast data, and can often help to even out potential bias or loaded question results. By triangualting both with data type sets as well as sources, you are building a more robust approach.
Iterative data gathering
It's often helpful to think about how you design and implement data gathering in stages, that is, by gathering some initial data and then progressing according to what you have at first found out. This can mean choosing methods in a variety of ways to suit your first results, and even using different methods concurrently to establish which method gathers the most interesting or relevant data. Iterative question development, based on initial findings can be especially helpful in keeping things focused and relevant.
A note on theory: In the Social Sciences disciplines, knowing your research paradigm is important, and discussing the parameters of your research within the context of a paradigm is significant. This might be more the case than in the Sciences disciplines.
The research paradigm is part of your discussion on methodology, and will contain theory and background reasoning for your decisions on your research design, including how you obtain your data, your research instruments, and how you intend to analyse your data. Common paradigms you might find within social sciences are critical perspective, interpretive, social constructivist etc.
The nature of knowledge (epistomology and ontology) and of perceptions in understanding and interpeting the world are concepts which though apparently 'high-brow' are actually very relevant to academic research. Factors and scenarios in the world which are researched need to be understood, interpreted and analysed, and knowing more about what the researcher is basing their knowledge and interpretation on can help to validate or challenge the findings of their research. Kinash calls the paradigm that the piece of research sits within the 'buyer-beware' statement.
Dr Kinash says "...methods are the techniques or processes we use to conduct our research. The methodology is the discipline, or body of knowledge, that utilizes these methods". I think this is a good way of thinking about the difference as it's really common for people to think they are the same thing.
(See more about methodology in the download PDFs and the helpful links sections.)