Camino House

Blog

Blog posts from Camino House

Receive occasional updates from Camino House. (We will store your data securely & will never share your contact details. See full Privacy Notice)

7 steps to design a successful research study

Before designing a piece of research, the most important thing is to scope it well. If you haven’t read it already, our blog post here will help show you how.

Now you’re ready to design your approach, follow these 7 steps to help make sure your research is as effective as possible.

1.      Using a set of post-it notes or small pieces of paper, write down all of the questions that you think will help you to answer your research question (and any sub-questions) you have identified when scoping your research. It can be very helpful to do this alongside somebody else or a group of people, so that you generate lots of ideas.

2.      Once you think you have a big pool of questions, lay them all out alongside each other, and group them into themes, where you think there are similarities between the question.

For example: asking ‘what events should we consider starting?’ is similar to ‘what events do you enjoy least?’, and also similar to ‘should we do more events?’ Depending on the scope of your research, you might group all of these under ‘Events’.

3.      Now you have divided your questions by theme, review each theme in turn:

a.      Are there any big themes missing that might help you answer your research question? If so, add this theme and jot down any questions which might help address it.

b.      Will the questions you’ve generated give you a good understanding of this theme? If not, add any questions that are missing.

c.       What order makes most sense when you’re thinking about how somebody else might answer these themes? You can think of this like a conversation… how would a natural conversation move from theme to theme?

4.      Once you’ve got all of your questions visible and thematised, it’s time for the big chop!

It’s the perennial problem of the researcher that there is so much that would be good to know that questions can multiply out of all control!

So, for each theme, look carefully at the questions you have suggested, and consider ranking them in order of usefulness. The criteria for top ranking should be that it best fulfils the ‘so what?’ test! In other words, if you find out the response to this question, what difference would it make to answering your research question (and ultimately, shaping your activity) Sometimes, this means de-prioritising questions that you think are more interesting in favour of questions that are more useful.

For example: Jenny’s overall research question is ‘How can we encourage more parents at school to join the Parent-Teacher Association’. Whilst she’s personally most interested in asking ‘What topics would you most like the PTA to discuss at our meetings’, she recognises that a more useful question to ask is ‘When would be the most convenient time for the PTA to hold our meetings’?

5.      When you have a ranked set of questions per theme, it’s time to put on your participant-perspective.

Thinking back to your scope, who will be participating in your research? And in what way?

With your participant in mind, decide how many questions you can feasibly ask for each area without putting them off. A good rule of thumb for a survey is that you don’t want it to take more than 10 minutes to complete. For a face-to-face conversation, consider the pressures of time on your interviewee and also on yourself / other interviewers. Any more than 30 minutes would be a substantial interview. For each project, the combination of time factors and the style of questions and participants will impact exactly how many questions you should ask, but it’s a good idea to pilot your survey / discussion guide before beginning the project properly.

6.      The second element of participant-perspective is to consider how you should ask the questions you’ve prioritised. Try rewriting them, remembering to:-

a.      Keep questions as short and understandable as possible

b.      When constructing questions, keep them balanced and open – don’t force a participant to answer in a particular way by only offering a narrow range of optional answers, or by phrasing the question to assume they agree / disagree

c.      Don’t assume that your participant will understand anything – keep questions jargon-free and explain difficult concepts or phrases where necessary

d.      Don’t refer to other questions when asking a new question – somebody should be able to read or hear each question you ask as a standalone sentence.

7.      Finally, once you feel you have designed a constrained set of questions which are in relevant, balanced language, and which help you answer your research question, you might finally think it important to include some ‘demographic’ information as part of your research questions. Confirming, for example, the gender of participants, their age (within bands) or their length of service with your organisation etc might help you to identify trends or similarities between people who share similar backgrounds.

You should only ask demographic questions which you think are relevant to understanding your particular situation, and for answering your research questions. Some people may find these questions intrusive, so you should also give them the opportunity not to answer if they prefer. This is especially the case with anything of a particularly personal character, including sexual orientation, religious identity, income etc.

 

Once you’ve designed your project, you’re ready to go. Visit caminohouse.com again soon to discover more about interpreting data for decision-making.

As with all data collection, there are standards for ensuring data is willingly offered and kept / accessed appropriately. If you are using an online research service (e.g. Survey Monkey), they will handle the legalities appropriately. If you are conducting more substantial research (especially if you are looking to make the research publicly available),  this guide provides a good overview.