Suppose you want to research a target group of as many as 10,000 people. Do you have to interview all 10,000 people to get the right results? Certainly not; only a part of the research population needs to participate in your research to get representative results. Let me explain what representativeness means and when the results are deemed characteristic.
Representativeness means the degree to which the respondents in a sample group are a good reflection of the target group of your research. Therefore, your research is representative, which means that the conclusion of your research is true for 'everyone' in your research population.
A sample
If you have a research population of 10,000 people, you will ultimately need to interview 400 people to arrive at the opinion of the larger group. This does not mean that you only need to approach 400 people. You have to deal with a response rate. This is the percentage of people who participate in your survey. Your response rate depends on the subject you are researching, how easy and fun it is to participate in your research, and what people get in return. I often use a response rate of 30% because I usually research fun subjects, and I am experienced in making it easy to participate in a survey. I also ask the client for a nice gift for the people who participate in the survey. Tips to increase your response rate
Because not everyone will participate in your survey, you will need to have a larger sample group. If you need 400 respondents and assume a response rate of 30%, you will need to interview a sample group of 400/30%=1,333.
When you conduct a survey, you must keep in mind that the smaller the research population is, the larger the number of respondents will be to arrive at the desired representative results. Sometimes, however, respondents' input is more valuable than the number of participants. In this case, you're referring to qualitative research, and it can be more important to focus on the research results than the representativeness in some cases.
If you want to know how many respondents you need for your research population: go to a sample calculator. These will often immediately tell you how many people you need to approach in your sample.
The range of research methods on offer is enormous, so you can sometimes not see the wood for the trees. That's why I give you a handy overview of different research methods in this blog. With some of the research methods, you will find a link to a blog with more information.
In this overview, I characterize quantitative and qualitative research methods. Would you like to know more about these? In my previous blog, 'When to choose: qualitative research or quantitative research,' you will find information about the difference between these two types of research.
Qualitative Research
- Group discussion: Conversation or discussion with several people about one or more topics.
- Individual interview: Structured or unstructured conversation with one person in which you go into detail about one or more topics.
- Literature Review: Research in which you use research, theories, and information already available (e.g., from a library or on the Internet) based on a problem statement.
- Mystery visitor: Research method in which you use experts who behave as customers or visitors and assess the quality of service or organization.
Quantitative research
- Observing: Observing actual behavior and recording responses.
- Questionnaires: Recording data and opinions of groups of people using a pre-prepared questionnaire. You can have this questionnaire completed digitally or in writing by a large group of people.
- Tear tickets: A research method in which you get a large group of people to answer one question quickly. This can, for example, be done by giving the audience a piece of paper with a statement on it before a show. After the show, the audience can indicate whether they agree or disagree with the statement by making a tear in the piece of paper.
- Existing source research: conducting research using existing datasets of quantitative data that other researchers have already collected. You then use the dataset again to answer a new question.
- Informal conversations: During an informal conversation that is already taking place, you will ask a few specific questions. You record the answers afterward and repeat them to multiple respondents. The respondents are not aware that they are participating in a survey.
- Ten-minute interviews: Short interviews to find out about respondents' experiences, opinions, and motivations. It is stated in advance how long the interview will last. You can use a timer.
- Tracking: following respondents (e.g., visitors to a museum or customers in a store) through a distinct area. This can be done through the Wi-Fi or Bluetooth of their own devices, but also with a device that you give to respondents, with which you follow them.
- Logbook: A document (digital or written) in which you have visitors or participants record events and specific data.
Observation is a method of finding out and recording actual behavior. You choose the observation method to determine how activities, programs, or projects are received and when you want to register spontaneous reactions. While observing, you make an objective report of what you see; there is no direct contact with the respondents. I give you some tips on how to observe effectively:
- When observing, use an observation list. With a structured observation list, you consistently record objective information, and you can quickly analyze the collected data.
- While observing, show an open and inquisitive attitude. Avoid jumping to conclusions about behaviors. Make an objective report of what you see.
- Name or record verbatim what you actually see happening, not what you think is happening.
- Don't just look at random gestures. A loose gesticulation doesn't say much. Its meaning becomes understandable only when combined with other gestures and signals.
- Depending on the situation, make it known that you will observe people, for example, if you will follow them during the activity.
- Combine observations with a (short) interview to indicate what you have seen.
- By making many observations, you can make representative statements about your research topic or group with supporting figures.
- The observers should be given clear instructions so that everyone records the observed in the same way and that interpretation of observation is avoided. To make different observers equal, you can do the first observations together.
- When observing, you can use all kinds of technical aids. There are various tracking devices on the market, but you can also consider a stopwatch to help you measure how long someone is using something.
- Always observe and consider the circumstances. If someone is standing in a cold room with their arms tightly crossed, chances are they are not defensive and closed, but simply cold.
- Make a schedule of when to observe in advance and do so at different times under different conditions so that the picture you get is as varied and complete as possible.
- Disrupt the situation you are observing as little as possible. Thus, do not talk to those being observed and be as inconspicuous as possible.
There are different things to consider when organizing and conducting a group interview than during an individual interview. With group interviews, you are dealing with group dynamics. Below are 10 tips for leading a group interview:
- Start organizing the group discussions well in advance. You otherwise run the risk that not enough people can join. Scheduling the group interview well ahead of time increases the likelihood that everyone will attend.
- Make sure the minutes are written by someone who is experienced in it. Writing the minutes for a group discussion is more complicated because of the dynamics of several participants speaking.
- If you are not sure the reporter will be able to keep up with the live coverage, then make a video or audio recording of the interview and write out the minutes later. Ensure that the recording equipment is present and ready to use before you decide to use the equipment.
- Make sure your interlocutors feel comfortable. Provide them with a pleasant ambiance, something to drink, and a snack.
- Have everyone at the table introduce themselves so that everyone knows with whom they are sitting at the table. Create a cheat sheet with names and functions/backgrounds for the participants at the table.
- Make sure everyone gets a chance to speak.
- Pay attention to the body language of the participants.
- Monitor time and stick to the agreed time frames. It's disappointing when you don't get to ask all of your questions.
- Explain what will be done with the results of the interview. Also, explain what you will do with the minutes made of the exchange (including the audio recording). Will you be the only one who has access to these files, or will you publish it? In the latter case, make sure you get feedback on the minutes before publishing them.
- Summarize during the interview and relay your statement back to the group to make sure everybody understood the message the same way.
Tweets
10 tips for holding a group discussion http://bit.ly/1L0zJW6
When having a group discussion, make a map of the table with names and function as a cheat sheet http://bit.ly/1L0zJW6
During a group discussion, make sure that everyone gets a chance to speak http://bit.ly/1L0zJW6
Pay attention to the body language of the participants during the group discussion http://bit.ly/1L0zJW6
LinkedIn
Government
Cultural Policy
Engineering policy
In my previous blogs, I explained how to determine the right research question and how to choose the proper research method. The next step in the research process is data collection. While gathering data, there are several things you need to keep in mind. I'll give you some tips:
- Determine your research population. It is not always necessary to interview everyone to get a representative picture. Especially in the case of large numbers, such as the voting behavior of all Dutch citizens, a representative sample is drawn.
- Keep a close eye on representativeness while collecting the data (in quantitative research). Does everyone in your sampling actually cooperate? If your subgroups are not equally represented, you can still question the missing respondents or decide to leave out a particular subgroup.
- Check whether you have enough information, especially in qualitative research. If this is not the case, you can conduct an additional interview.
- The data must be registered in a way that the data is clearly arranged and ready for analysis. Having a large pile of completed questionnaires is in itself insufficient for analysis. The best way to register results depends on the research method you have used.
If you are going to conduct research, it is essential to formulate the right research question. The purpose of the research and the research question indicates what the setup of the research will contain. A good research question ensures that you have the correct information to actually work at the end of the research. There are five steps to formulating the right research question:
Step 1: Why research?
The first step in deciding on a proper research question is to answer the question, "why are you going to research, evaluate, or monitor? Answering these questions will give you an idea of the type of research question you will ask, what information you need, and which answers you need. Do you want to improve a project, or do you want to justify your project? With both goals, you're going to be evaluating, but the answers you wish to obtain afterward are very different—improvements versus demonstrated effects.
Step 2: What information do you need?
Based on your goal (Step 1), you can determine what information you need to meet this goal. What do you need to know to meet your goal? Do you want to improve goal attainment? If so, you need information in the form of points for improvement (what is going well and what could be improved?) Is the goal to legitimize your project for funders? Then you need information in the form of the impact of the project. Do you want better alignment with the target group, and is this the goal of your research? Then you need information in the form of characteristics, wishes, and needs of the target group. Always ask yourself the question: what am I going to do with this information? This will enable you to set priorities.
Step 3: Formulating a research question
Based on your research goal and the accurate description of your information needs, you can formulate your research questions. Formulate these questions as precisely as possible. When preparing the questions, remember that you or a researcher will soon be answering them. So will you achieve your research goal, and can you take follow-up action if you get answers to these questions?
Step 4: Preconditions
Various preconditions can be linked to the research. It is essential to gain insight into this in the phase of formulating the research question. Preconditions you need to consider include time, budget, availability of specific data, and so on.
Step 5: What will the final result look like?
Research results can be delivered in a variety of forms. <LINK to blog>Digitally, physically, a report, a fact sheet, a book, a video, an article, a discussion session, a presentation, a website, a PowerPoint presentation, an infographic, an animation, a magazine, a flyer, to name a few. And what is communicated therein? Impacts, arguments, or areas for improvement. Make a representation of what you would like to get at the end of the research and whether you can use it to take your intended action, convince others, improve policies or activities or make a decision.
A questionnaire is used as a measuring tool to answer your research questions. It is essential that you can use the results of the questionnaire. The quality of the questionnaire determines the quality of the data collected. It is therefore important to carefully formulate the survey questions. Here are a few tips and rules of thumb to help you formulate well-structured questions:
- Keep the questions simple. Don't use complex language, and consider your audience in your choice of words. For example, "What do you prefer to do in your free time?" is better than "How do you prefer to relax?
- Provide an unambiguous interpretation. 'I am satisfied with the quality of the exhibition' is an example of how not to do it. What is quality? The type of artwork, interactive elements, crowds, venue, the light?
- Formulate the question as precisely as possible. Refer to place and time and mention numbers. Try to delineate questions such as "Have you recently..." to a specific period, for example, "Have you in the past six months...".
- Avoid vague wording and avoid terms like ‘often’ and ‘sometimes’, also in the answer categories. Everyone interprets often and sometimes differently, so it is better to ask for a specific number of times.
- Avoid duplicate questions. No 'and' or 'or' in the questions. A question like "What did you think of the performance and the actors?" cannot be answered with one answer if the audience thought the performance was a little off, but the actors were excellent.
- Do not formulate (double) denials in the question. A negation in the question is confusing. For example, 'I am not dissatisfied with what I have seen' or 'I don't like to visit a museum.'
- Ask short questions.
- Be careful with examples in the question and suggestive questions. Chances are that the respondent may only think of these examples. This can happen with a question like 'How often do you undertake a cultural activity such as a visit to a museum or a play.' The respondent will be inclined to think that a visit to a festival or a dance performance is not part of the equation.
- Make sure the question measures up: The question should answer the research question. If you want to know if an exhibition inspired someone, do not ask how long they stayed. It is possible that the length of stay was longer or shorter because the respondent had to wait or had to leave earlier and had no choice.
- The answer categories to the questions should be mutually exclusive, and it should be clear to the respondents which answer to tick/indicate.
- Provide the same direction in the response order for scale questions. If at one point you are asked to rate something on a scale from totally disagree to totally agree (increasingly positive) and a few questions later on a scale from very satisfied to very dissatisfied (increasingly negative), there is a chance that people will fill this out incorrectly.
- For scale questions, try to keep the scale the same for each question. So do not use a scale of 1 to 5 for one question and a scale of 1 to 7 for another. This also makes it easier to analyze.
- See which scale fits best. A rating scale of 1 to 10 offers a lot of variation but is also more challenging to interpret (for some, a 10 is good, others think a 10 is perfect and therefore don't give it easily). A 2 (good/bad) or 3-point scale (good, average, bad) offers minimal variation and makes it harder to answer a question if it is an opinion (opinions are typically nuanced). An even scale causes a respondent to have to choose; with an odd scale, you offer the opportunity to sit safely in the middle.
- Also, provide the option of a reasoned non-answer, for instance, by creating an answer option like not applicable or don't know/no opinion.
It is not always necessary to speak to everyone to get a representative picture. Especially with large numbers, such as the voting behavior of all Dutch citizens, a representative sample is drawn. A representative sample contains a sufficient number of respondents. Be aware of drop-outs; not everyone in the selection will cooperate with your survey, and all subgroups should be represented, men-women, young-old, ...
Two methods
You can draw your sample from a database of addresses (and background characteristics), such as ticket ordering service (theater) or a purchased address file. You can also ask people directly for your visit, for instance, at the entrance/exit of a museum/theatre. In this case, you determine in an objective way who is asked, for example, by asking every 5th person.
Three steps
- To ensure that your sample is large enough, you first need to know the minimum number of people who should participate in your study. This depends on the size of your research population, the reliability you are aiming for, and the margin of error you are willing to accept. Calculating your required sample size is done with a sample calculator.
- Next, you need to assess the likelihood of people cooperating with your research. People on the street are less likely to cooperate with your research than if they have an affinity with the subject (e.g., the theater they frequent) or if the questionnaire is done in a classroom setting. Based on this, you will determine how many people you should ask.
- Then you're going to see if you have any subgroups in your research population. You can take this into account when drawing the sample or check and correct it afterward. If you take this into account beforehand, you will draw a selective sample. Consider which subgroups you want to take into account and how large these groups are. This is especially easy if you have a database with addresses and characteristics. But also, if you approach respondents directly (e.g., at the entrance of the museum), you can take this into account (only ask men or mjk holders).
Afterward, when you check for proper representation of the subgroups, you draw an a-selective sample. With the completed questionnaires, you're going to check that all subgroups are equally represented.
Individual and group interviews allow you to find out a lot. For example, about the impact of an activity, the course of a program, or the influences of different actors.
But how do you get the most out of your interviews?
- Make sure you are talking to the right people. Thoroughly explore your options in this respect. Who can answer the questions you have? What perspectives are out there?
- See if interlocutors can complement each other (group interviews) or tell more when they are alone (individual interview). When conducting a group interview, make sure that everyone gets a chance to speak.
- Make sure your interlocutor(s) feel at ease. Provide a pleasant space and beverages. Provide anonymity if necessary. Be polite and respectful to your interlocutor.
- Make sure your questions are consistent with your research topic and your interview questions are all-inclusive.
- Prepare your interview by formulating questions that are attuned to what your interlocutors know.
- In an introductory narrative, tell who you are, why you are conducting this research, your research question, who your client is, whether it is anonymous or not, and what the interview is essentially about.
- Formulate your questions in a way that your interlocutor understands them.
- Stay objective. Keep your opinions to yourself.
- Let your interlocutors finish speaking. Three seconds of silence does not mean that they finished speaking. They may want to think about the rest of the answer.
- Pay attention to your interlocutor's body language.
- Try to make the interview as natural as possible. Don't read the questions from your question list. Make a connection with what has already been said. Don't interrupt your interlocutor only to return to that topic later. At the end of the interview, check whether you have asked all the questions.
- Summarize your interlocutors’ answers so that they know you have understood them correctly and so that they hear the mainline repeated.
- Appeal to your interlocutor's responses.
- Monitor the time and adhere to the agreed-upon time. It is frustrating if you have not been able to ask some of your questions because your interlocutors have to go to their next appointment.
- Explain what happens to the conclusions of the interview. Also, explain what you will do with the report made of the interview (e.g., the audio recording). Will you use it only to refer to it yourself, or will you also publish it or hand it over to your client? In the latter two cases, make sure you give everyone feedback on the report.