4.6 Forms of Primary Research

Forms of Primary Research

Primary Research versus Secondary Research

Most students are familiar with secondary research, even if they don’t know what it is called. Secondary research is the kind of research you do in the library or online. When one is conducting secondary research, they are looking for sources of information that other experts, writers, and thinkers wrote about a subject. We call this kind of research ‘secondary’ because it relies on others to have collected the research and written about it.

This is in contrast to primary research. Primary research is conducted, not in the library or online, but in the world. When conducting primary research, a researcher will use one or more tools, or methods, to collect data directly from people or the things they are studying rather than from books or texts already written about those things or people. One example of this is the kind of research a journalist does. A journalist may look up information that has been written about the news they are investigating, but they will also go out and talk directly to other people about that news–seeking out actual witnesses on the scene or officials in charge. In the job of a journalist, this is especially important since their subject, the news, involves late breaking events that may not have been written about yet.

Note that although primary research and secondary research constitute different kinds of data collection, they go hand in hand. Most researchers will conduct both secondary research, collecting relevant information on their subject that has already been written and published, and primary research, collecting new data and evidence that no one else has collected before.

Primary Research Methods

Popular culture is rife with images of the solitary scientist locked up in her laboratory, combining the contents of test tubes or prodding a lab rat through a maze. Usually, such a scientist is engaged in one method of primary research called experimentation, in which a researcher will set up a series of tests or demonstrations in the controlled setting of a lab in order to test his or her hypothesis. What isn’t made evident in popular culture is that scientists, scholars, and researchers can actually choose to engage in a variety of different forms of primary research, depending on their field of study and the kind of knowledge they want to discover. Other examples of primary research methods include observation, interviews, focus groups or panels, surveys, and ethnography. In this class, you will probably only conduct primary research using a couple of these methods. But learning about each of them will give you a better understanding of the kinds of research that scholars and experts might do. And since scholars and researchers also write to report their primary research, it will also help you better understand the studies, reports, and articles you find when you do secondary research.

Below some common methods of primary research are defined:


This method involves going out in the world and watching, using your five senses to collect data. Observation as a primary research method occurs wherever a person establishes a specific set of criteria, features, or characteristics and then examines something for those characteristics or features.

Here are other ways observation might be used:

  • One might observe a group or organization, exploring how business is conducted or how people in the group communicate
  • One might observe artwork or other man-made or natural objects in order to interpret that artwork.
  • One might view and record observations from several people’s Facebook pages to examine how this kind of social media commonly gets used.
  • One might observe memorial spaces in public parks at various times in a day to record how the public makes use of those spaces.
  • One might review or conduct an analysis of a text or film, judging it based on certain aesthetically appealing qualities or characteristics.

Observation is great for inquiry in which you either can’t ask questions (for instance, a monument or painting won’t talk back) or because you want to collect information on how something works without interfering by participating yourself or asking questions for which you may or may not receive the best answers.  At the same time, observation means you can only observe one or a few examples, thus it is hard to say that anything you observed is true for most or all situations.

Planning Observational Research

The plans a researcher makes to conduct observation depends on he kind of data she or he wants to collect. Observational data can be qualitative or quantitative.

For qualitative data, you describe, in words, what you see, taste, touch, hear, and/or smell as well as what these observations may mean. One common way to do such qualitative observational research is to use the DIE method – Describe, Interpret, Evaluate.

  1. Describe means that you watch, using your five senses, and write down everything you see, taste, feel, hear, or smell.  You do not make any assumptions or come to any conclusions about what you are describing, but simply record what is physically observed.
  2. You interpret only afterward, figuring out after careful observation and description who, what, where, when, why, and how based on what you have observed.
  3. After that, you can evaluate, determining how the things you observed relate to the ideas you are investigating and, in particular, to your research question(s).

This strategy allows you to carefully distinguish between what is actually being observed, the reasons the phenomena you are observing are happening, and what you think it all means.

Another strategy to conduct qualitative observational research, which can be used alone or in conjunction with the DIE method is a dual-entry notebook.  On each page of a dual-entry notebook, you create two parallel columns.  In the left column, you describe what you observe (remember the five senses), and in the right column, you analyze and interpret what that descriptive data might mean.  This allows you to jump back and forth between description and analysis, while distinguishing between what is actually being observed and what the significance of it all might be.

With either of these qualitative methods, it is good to have a list of qualities or characteristics you expect to observe so you are prepared for what to watch for and examine. This list of attributes may change as you discover new things about your observed subject, but such a list helps keep your observations focused.

Quantitative observation usually involves tallying – simply making a mark every time the phenomena you are observing happens. This allows you to calculate the frequency or number of anything being observed. To do this, you must select periods of time in which to collect data and decide beforehand a certain set behaviors or phenomena you will count during each observation period. After that, of course, you must observe and tally those behaviors or phenomena.  After you collect these numerical results you can interpret the data and evaluate it in terms of your research question(s).

Whichever kind of observation you perform, it will require you to make a plan to decide what kinds of things you will look for when you observe (what kind of phenomena fits the bill for the research question you’re trying to answer). It will also often require that you plan certain times and/or place in which to do your observations. This is especially the case when you plan to observe things that happen a different times or day and/or in various locations.

When doing observational research, it can often be useful to record what you are observing, either photographing or video-recording it. This is useful because it allows you to look at it again and again. Keep in mind, though, that if a researcher records people in a way that would make them identifiable by others, they must gain permission to use those images or footage from the individuals recorded.


Interviews involve one-on-one sessions with individuals, in which you ask open-ended questions. You collect their broad, open-ended answers much like you do with observation, without coming to conclusions or assumptions about what the person is saying. Only afterward do you analyze the questions and relate it to the subject and your research question(s).

Interviews come in a couple different varieties. One version is a representative interview, in which you interview people who are affected by or experience a certain problem or issue. Another version is an expert interview, in which you interview people who are experts, scholars, professors, or professionals in a field related to your topic of research.

It is even possible to combine interviews with observation, by asking interview respondents to view something (like a video or a set of images) and then ask questions about what they think or noticed about the phenomena or artifacts they viewed.

Interviews are quintessentially qualitative, leading to complex understandings and viewpoints of one or a small group of people. Generally, the answers are in depth and nuanced because the respondent has some time to construct his or her answers carefully and add clarification if needed. Another good thing about interviews is that they allow you develop specific questions tailored to the individuals you are interviewing and to change your questions or come up with new questions based on the respondent’s previous answers. In this way, the interview becomes sort of a conversation; the information you collect adjusts and changes according to what you discover at that moment.  Other the other hand, interviewing limits the number of people from whom you can collect information, so it isn’t as good for coming to conclusions about what most or all people think.

Planning Interview Research

You can conduct interviews in person or via writing (email, chat, instant message, etc.). Interview data can be collected via the interview respondent’s writing, by written notes you take as the interviewer, or sound-recording or video-recording. Part of planning an interview requires you to decide how you’ll conduct the interview and how you’ll collect the answers.

In general, an interview is conducted face-to-face or live over the phone. This approach to interviewing allows you to adjust questions and come up with new ones as the conversation proceeds and new ideas emerge. It can be challenging when it comes to collecting data, though. Will you take notes to record the respondent’s answers? If so, you may need to halt the interview at times to get the information recorded or ask the respondent to confirm what they said. Will you record the interview? In this case, you must get permission from the respondent to do so and you must take time to listen to the interview later in order to take notes on it. Something similar can also be done via a teleconferencing tool like Skype or even through instant messaging.

Interviews can also be conducted via email, where you send the set of questions to the respondent and they answer them on their own time. This eliminates the need to plan a time to meet and talk to the respondent. It might also lead to longer, more in-depth answers since the respondent will have the time to think about and write down their responses. What’s more, once the respondent replies, all the data is already collected in the email they send back. On the other hand, an email interview does not allow an interviewer to rephrase questions or add new ones. If a respondent didn’t understand a question, there isn’t a way to rephrase to get a better answer. Also, when conducting an email interview, it is helpful indicate or negotiate with your respondent a deadline by which you’d like to receive answers from them. This way, they know when they need to plan to sit down and write answers for your questions and you increase the likelihood that you receive those questions when you need them.

To conduct interviews of multiple respondents, you also will need to come up with a common set of questions to ask each person so that the same kind of information can be collected and compared. These questions must be phrased in such a way that they are understandable to the people you are interviewing. In addition, you must plan ahead by contacting participants to interview, set up times and places to meet them, etc.


Surveys involve developing a series of short, easy to answer, multiple choice or multiple answer questions that are distributed to a large number of people. Usually, surveys are used to collect quantitative data; a researcher will total up each kind of answer for each question and calculate the mean (average), median (middle), and mode (most common) of those answers. As well, other statistical analysis can be done on survey data to mathematically determine how significant or remarkable certain answers are. In any case, the numerical data collected from a survey is then interpreted, looking for answers they provide to research question(s).

Surveys are great for collecting information about large groups of people, since you can distribute surveys widely, collect them as a group, quickly total up answers, and do calculations.  Because of this, you can begin to make conclusions based on how representative your survey sample is of the larger group you are investigating.  A good survey sample means that you can assume that even people you didn’t survey will likely answer in the same way as those you did survey.

On the other hand, you cannot collect very complex information through a survey since the people who take the survey are automatically limited in the kinds of answers they can give and the questions and answers have to remain general enough to refer to and be understandable by all people.

Planning Survey Research

Surveys will require a set of survey questions applicable to the research question(s), identification of a sample population, a way to get surveys out to that sample population, and time to collect returned surveys and calculate the data.

It can actually be quite tricky to phrase the questions and answers in a survey effectively. Since all the question need to be understandable to a large number of people, a researcher must be careful to phrase the questions in simple terms and provide explanations for more complex ideas or terms that respondents might not understand. Since most questions in a survey are multiple choice, researcher also needs to be sure they have provided all the possible answers respondents may want to give to each question, or even have an “Other” option in case the answer a respondent might give isn’t listed. Some of the questions asked might be demographic questions – questions about respondents’ age, gender, race,political or religious affiliation. These questions help a researcher determine whether the people they are surveying matches they population they want to collect data on.

There are various ways to distribute surveys: a paper form that is passed out and collected, an in-person survey in which questioned are asked and responses are collected face-to-face, or an online form that is distributed via email or some other digital media. If using online tools to write and distribute a survey, you can use tools like freeonlinesurveys.com or surveymonkey.com.

Primary Research Methods: Strengths, Planning, & Ethics

Below is a table breaking down on each method of primary research and their major features.

Observation Interview Survey
Definition “Gathering data through your own senses” (Miller-Cochren & Rodrigo,  p. 93) “Asking questions of one or more people in person” (Miller-Cochren & Rodrigo, p. 93) “Asking [short answer] questions of large groups of people” to generalize public opinion (Miller-Cochren & Rodrigo, p. 93)
  • Permits collection of data regarding subject matter people may not discuss honestly
  • Permits research about things and direct observation of behavior rather than of people and/or their assumptions about behavior
  • Does not need to be coordinated with specific human beings
  • Provides in depth answers/ opinions
  • Provides  opportunity for clarification / elaboration
  • Adaptive – questions can change based on how interview progresses
  • Comparative – can compare answers among respondents and look for similarities/ differences based on respondent characteristics
  • May generalize about larger populations
  • Generally easy to provide answers
  • Large collection of quantitative data
  • Comparative – may be able to compare answers among different respondent groups (men vs. women, rich vs. poor, student vs. teacher, etc.)
  • Easy to offer charts or diagrams based on answers
  • Quantitative (Tallying)
  • Qualitative (Describe artifacts, behavior, texts, images, etc.)
  • Representative
  • Expert
  • Multiple Choice
  • Multiple Answer
  • Likert Scale (Strong Agree, Agree, N/A, Disagree, Strongly Disagree)
  • Rating/Ranking
  • Combo
  1. Identify objects or contexts of study.
  2. Conceive characteristics or behaviors to observe based on research question(s)
  3. Prepare a method of recording data
  4. Set up times/places for observation
  5. Record data
  1. Identify respondents
  2. Plan & schedule method and time/place of interviews
  3. Write open-ended questions relevant to research question(s)
  4. Prepare method to collect answers
  5. Conduct interviews
  1. Identify a population
  2. Write questions based on population and research question(s)
  3. Prepare (and test) survey questions and answers
  4. Distribute surveys to population
  5. Collect surveys
  • Double entry notebook
  • DIE (Describe, Interpret, Evaluate)
  • Table/chart (Tallying)
  • Face-to-Face
  • Phone/Skype
  • Email
  • Recording by printed/typed notes, voice recording, email
  • Print survey
  • Online survey
  • Live survey (face-to-face or p/hone)
  • Oberserver’s paradox – the presence of a researcher can skew results, change people’s behavior, etc.
  • Observations must be conducted honestly and systematically.
  • Permission must be obtained to record people’s faces or voices.
  • Etiquette
  • Respondents must be told the purpose and use of interview results.
  • Permission needed to use answers and names of people.
  • May need to confirm that anonymity will be maintained.
  • Permission needed to record voice / use voice.
  • Answers need to be recorded honestly and systematically
  • Age – If respondents are children, the permission of parents are needed.
  • Sample population must be representative of the whole.
  • Respondents must be told the purpose and use of survey results.
  • Permission needed to use answers.
  • Must confirm that anonymity will be maintained.
  • Age – If respondents are children, the permission of parents are needed.
  • Unbiased / non-leading questions must be used.
  • Must provide all possible choices for honest answers
  • All surveys collected must be used.

Works Cited

Miller-Cochran, Susan, and Rochelle Rodrigo. The Wadsworth Guide to Research. Cengage Learning, 2009.