Skip to main content

Primary Data Collection

How do I gather primary data?

In the social sciences (such as psychology, communication and business) and humanities (such as history and English), there are a number of different ways to collect primary data. Remember, if you are working with living people and/or collecting sensitive information, you are required to submit your research plan for an ethics review. By undergoing an ethics review, you make sure that you are not violating any laws and you are promising to treat the participants and their information with respect and care.

Key informants

Using a formalized process, a well-trained interviewer asks predetermined questions in a face-to-face setting.

  • Data are collected through in-depth interviews. The answers to the questions will form the data for the research.
  • Key informants are experts in their respective fields; thus, they are purposefully selected based on the needs of the research. For example, Thomas Homer-Dixon may be interviewed for insights on how science and technology has affected society and our environment.
  • Interviews are usually conducted in a comfortable location such as the interviewee's home or office.
  • The interviewer uses probing questions designed to avoid yes or no responses.

Focus groups

In a group setting, a moderator will guide an informal discussion about a research topic. The group normally consists of between 6 to 12 individuals and the moderator (usually the researcher). The moderator will:

  • Guide the discussion;
  • Ensure that the key aspects of the topic are discussed;
  • Observe the reactions of the participants; and
  • Record the data (using audio files, video files, field notes, etc.)

Case studies

A case study is an in-depth examination of a single unit of interest. The researcher may select one person from a focus group to gather more detailed information from that person. Alternatively, the researcher may focus on one company or business to examine how that company works, rather than looking at a whole industry. For example, rather than examining how and why the automotive industry has changed over the last 20 years, a researcher may focus only on General Motors to determine why it (or some part of the large corporation) failed (or succeeded). The data generated will be very detailed.


You must be a detail-oriented investigator and able to see patterns in the data.


The researcher quietly observes subjects in a particular situation. The researcher looks for how the participants act and react under certain situations. This is a useful technique for assessing behaviours. An observation study may be conducted by:

  • Human (researcher) or mechanical (video, photos, etc.) observation
  • Contrived observation (i.e. staging a situation)
  • Naturalistic observation (i.e. observing subjects in a natural setting when the subject does not know he or she is being observed)
  • Disguised observation (individual is unaware of observation, e.g. security camera at a grocery store) or non-disguised (individual is aware of observation, e.g. actors in a TV show).


Formal questionnaires

Well-written questionnaires are difficult to create. Questions may have yes/no answers (i.e. closed- ended questions) or answers on a scale. Closed questions allow for a limited choice of answers. Scales allow for more possibilities, but still do not allow participants to elaborate on their answers. Open-ended questions may also be used depending on the nature of the research. Open-ended questions do not allow for yes/no answers; they normally encourage participants to express their thoughts in their own words which allows for a more complete picture.

Closed questions
  • Do you get along well with your parents?
Scale questions

Scaled responses measure degrees of agreement or disagreement. For example:

  1. How would you rate the service at our restaurant?
    1. Very poor
    2. Poor
    3. Average
    4. Good
    5. Very good

An advantage to this type of question (and to yes/no questions) is that they allow you to use various electronic and non-electronic methods of distribution, collection, and analysis. These tools are often available for little or no cost, making surveys relatively inexpensive to conduct.

Open questions
  • How well do you get along with your parents? (Or, Tell me about your relationship with your parents.)

Open-ended questions require more analysis and interpretation. They can be more time-consuming for both the participant and the researcher.


  • Care must be taken in crafting questions to ensure honest and unbiased answers. Make sure that your questions are worded so that they do not lead the respondent to a particular answer.
  • Before starting your research, check with your professor or TA to help decide if primary research is needed, if your plan is appropriate, and if you also need secondary sources.
  • Even if you conduct primary research, you usually also need to conduct some secondary research to see what other people have discovered in your area of research.

How should I decide when to use primary data?

Primary research can be costly, difficult and time-consuming. Make sure that this type of research is important to your project and consider the following:

  • Do I have the budget to conduct the research?
  • How much time will it take me to collect the data? It usually takes longer than you expect to do any kind of primary research.
  • Are the data essential to the success of my project?
  • How can I ensure that my data are accurate? (i.e. how can avoid skewing my data?)
University of Ontario Institute of Technology logo