Overview of Primary Research

Much like how different academic fields have different Discourse Communities, how research is defined varies widely from field to field, and, as you progress through your college career, your coursework will teach you much more about what it means to be a researcher within your field. For example, engineers, who focus on applying scientific knowledge to develop designs, processes, and objects, conduct research using simulations, mathematical models, and a variety of tests to see how well their designs work. Sociologists conduct research using surveys, interviews, observations, and statistical analyses to better understand people, societies, and cultures. Graphic designers conduct research through locating images for reference for their artwork and engaging in background research on clients and companies to best serve their needs. Historians conduct research by examining archival materials—newspapers, journals, letters, and other surviving texts—and through conducting oral history interviews. Research is not limited to what has already been written or found at the library, also known as secondary research. Primary research is research that is collected firsthand rather than found in a book, database, or journal.

Primary research is often based on the principles of the scientific method, a theory of investigation first developed by John Stuart Mill in the 19th century in his book Philosophy of the Scientific Method. Although the application of the scientific method varies from field to field, the general principles of the scientific method allow researchers to learn more about the world through observable phenomena. Using the scientific method, researchers develop questions or hypotheses and then collect data on events, objects, or people that is measurable, observable, and replicable. The ultimate goal in conducting primary research is to learn about something new that can be confirmed by others and to eliminate our own biases in the process.

This section explores some common ways of conducting primary research:

• Surveys. Asking participants about their opinions and behaviors through a short questionnaire.
• Interviews. Asking participants questions in a one-on-one or small group setting.
• Observations. Observing and measuring the world around you, including observations of people and other measurable events.

How do you choose between a survey, an interview, or an observation? It depends on what kind of information you are looking for. You should use surveys if you want to learn about a general trend in people’s opinions, experiences, and behavior. Surveys are particularly useful to find small amounts of information from a wider selection of people in the hopes of making a general claim. Interviews are best used when you want to learn detailed information from a few specific people. Interviews are also particularly useful if you want to interview experts about their opinions. Observations are useful for gathering data about actual human behavior by recording it as it occurs. In sum, then, use surveys to learn general patterns from many people, interviews to gain details from a few people, and observations to determine how people behave or act.

Of course, there are other ways of conducting primary research that may be more common in your particular field of study:

• Case Study. In-depth analysis of a person or group of people over a period of time.
• Focus Group. Planned small-group discussions around a particular topic.
• Data/Text Analysis. Analysis of an existing collection of data or texts.
• Clinical Trials. Study of a medical approach, device, or treatment.

Surveys

Use a survey to collect general information from a large group of people. Surveys tend to be more quantitative than qualitative. This page discusses the benefits of surveys, the limitations, the types, and the “3 C’s”  when creating surveys.

Benefits

A survey is one of the more efficient ways of getting a large amount of data in a relatively short amount of time. If your research question is “What do [insert demographic] think/know about [insert idea, subject, event]?” then probably a survey is a great place to start. With the right questions and sample, you can get a good idea on how people of a certain population perceive the subject/idea you are researching.

Limitations

Any research method you choose will have limitations. A common limitation to surveys is the lack of “follow-up” questions. For example, if you have 50 students take your survey and it’s completely anonymous, it will be extremely difficult to follow-up with the subjects and ask more questions. On a related note, most surveys are multiple choice and/or short answer, so the subject only has so much “space” to express their opinions/thoughts on the matter. If you are wanting more in-depth opinions/thoughts on a subject, you will need to either format your survey to allow for that or do an interview instead.

Another limitation to consider is that your data will be self-reported, meaning that you have to rely on your subjects’ answers. Several factors could influence your subjects’ responses, either consciously or subconsciously, that may not be completely honest. Common reasons for unreliable data include your subjects wanting to portray themselves in a more positive light, your subjects may assume the “point” of the study and thus want the point to go their way even if their answers are not entirely honest, your subjects are embarrassed to reveal specifics about their life, or your subjects may not be aware of their own biases on certain subjects.

Question Types

Here are a few types of questions you can use:

Multiple-Choice Questions. This is a common type of question on surveys. Ask a question and/or give a statement, and then the respondent has to choose from a list of answers. If the answer is based on opinion or preference (e.g., favorite types of music), it is best to give an “Other” answer option and a fill in the blank space.

Matrix Questions. Along with multiple choice questions, there are matrix questions that use Likert-type scales. These are common in many social and applied science disciplines. Usually, the researcher gives one or more statements or questions (e.g., “I enjoy listening to Beethoven,” “I enjoy listening to Mozart,” etc.) and then the answers are in a Likert-type scale (e.g., Strongly Agree/Agree/Neither Agree Nor Disagree/Disagree/Strongly Disagree).

Short Answers. This is where qualitative and quantitative methods mix. The researcher gives either a closed-ended question (e.g., “Did you enjoy Star Wars?”) or an open-ended question (e.g. “Did you enjoy Star Wars? Why or why not?”) If the survey is on a piece of paper, the respondent has a finite space to answer. If the survey is online, most online surveys let the researcher choose the character limit for the text box response.

3-Cs: Clear, Consistent, & Concise

You want the survey and the responses to the survey to be clear, consistent, and concise. This is not only important for logistics (e.g., the subjects understanding the question, having clear data, etc.), but it is also important from an ethical standpoint. Whom you send the survey to, what types of questions you ask, and what kinds of answer-choices you give can skew your data one way or another, and knowingly doing this can be seen as unethical.

Be Clear.

Be as clear and straightforward as possible for your subjects. Clarity is not only helpful for your respondents but is also helpful for you as well in understanding your data. Below are examples of unclear questions.

Double-Barreled Questions. Avoid combination questions. The respondent may have a certain opinion for one part of the question, but a completely different opinion for the other. It is best to separate these, then, into two questions.

Example:

• Unclear: The food & service was great. Agree or Disagree?
• Clarified: The food was great. Agree or Disagree? The service was great. Agree or Disagree?

Double Negative Questions. While researchers may easily understand the questions they are asking, a double-negative question can be confusing for respondents. Try to avoid negative statements as much as possible.

Example:

• Unclear: I am not happy when my food is not hot. Agree or Disagree.
• Clarified: I am annoyed when my food is cold. Agree or Disagree.

Biased/Leading Questions/Answers. In this case, sometimes it might be too “clear” what the researcher wants. You must be aware of your own biases and make sure that you are not leading your respondent. When leading questions are put into surveys or interviews, the researcher, at best, is viewed as presumptuous; at worst, purposely misleading.

Example:

• Unclear: Why do you love this restaurant? Fun/Friendly/Delicious
• Clarified: What do you like about this restaurant? Fun/Friendly/Delicious/Other/Do not like this restaurant/No opinion

Ambiguous Terms. Be specific in your terms, both in the question and answer options. Choices such as “seldom,” “rarely,” “some of the time,” etc. can be interpreted differently depending on the subject.

Example:

• Unclear: Pick the answer that best describes how often you come here: Always/More than often/Often/Most of the time/Sometimes/Seldom/Rarely/Never.
• Clarified: Pick the answer that best describes how often you come here: Daily/Weekly/Monthly/Every few months/Once a Year/First time/Other.

Be Consistent.

Consistency not only helps your subjects understand the questions/answers but also will help you as a researcher when looking at your data and finding patterns.

Consistent Terms. While having clear terms, you need to also have consistent terms throughout.

Inconsistent Example:

Q#1 Have you injured your kneecap before?
Q#2 If so, when did you hurt your patella?

Consistent Example:

Q#1 Have you injured your patella (kneecap) before?
Q#2 If so, when did you hurt your patella?

Consistent Order of Answer Choices. Along with consistent terms, the order of the answers need to be consistent as well.

Inconsistent Example:

Q#1 Parking on campus is convenient.
Strongly Agree/Agree/Neither Agree Nor Disagree/Disagree/Strongly Disagree
Q#2 Parking downtown is convenient.
Strongly Disagree/Disagree/Neither Agree Nor Disagree/Agree/Strongly Agree

Consistent Example:

Q#1 Parking on campus is convenient.
Strongly Agree/Agree/Neither Agree Nor Disagree/Disagree/Strongly Disagree
Q#2 Parking downtown is convenient.
Strongly Agree/Agree/Neither Agree Nor Disagree/Disagree/Strongly Disagree

Be Concise.

From your questions to the survey as a whole, the more concise you can be, the clearer your survey will be to your respondents—which means you may even get more responses. Remember that your subjects are doing you a favor by taking time out of their day to take your survey.

Unnecessary/Irrelevant Questions. For example, if you want to know what students think about parking on campus, asking them what their major is might be unnecessary. Be direct and to the point. What do you want to know, what do you want to know from your subjects, and what questions help you get those answers?

Text-Heavy Questions. Sometimes giving subjects a hypothetical scenario might be needed, but if you have 10 questions that are more than a paragraph each, your subjects may “check-out.”

Survey Fatigue. This is a known phenomena where the subject becomes tired/bored (i.e. fatigued) with questions. Respondents will either not finish the quiz or will mark answers without even reading the questions.

Repetition. Don’t ask your subjects “Do you park on campus?” and then “How often do you park on campus?; only ask “How often do you park on campus?” with the answer option of “I don’t park on campus.” Respondents may get frustrated if they feel they’ve already answered the question.

Lengthy Surveys. Each survey/research question is different, so there is no “magic” number for how many questions your survey should have. However, if you are just asking the “general public” about an issue, shorter will more than likely be better. A 5-question survey about parking on campus will likely yield more data (and probably more applicable data) than a 30-question survey.

Interviews

Interviews, or question-and-answer sessions with one or more people, are an excellent way to gather in-depth information from a person or group of people for your primary research project. In contrast to surveys, which gather data from a large sample group of people, interviews can offer you the chance to gather more detailed information from one person or a small group of people. This page presents information on how to conduct a successful interview, including writing good questions, choosing the right people to interview, selecting an interviewing method, recording your interview, choosing an interview location, and transcribing your interview.

What to Ask

A good interview starts with good, clear, and unbiased questions. Generally, for an interview, you will want 5-8 open-ended questions that invite your respondents to explore your topic with some depth. Questions that start with “how” or “why” almost always allow for that kind of exploration, and “what” questions that do not have simple one-word or one-phrase responses can work quite well, too. The structure or organization of these questions matters, which means you want to keep questions around similar topics together so that your participants can answer your questions more easily and thoughtfully. Depending on your interview set-up, you might want to use a semi-structured interview approach, in which you ask follow-up or clarification questions based on a respondent’s answers.

Whom to Interview

One of the keys to a successful interview is choosing the right person to interview. Think about whom you would like to interview and whom you might know. Do not be afraid to ask people you do not know for interviews. When asking, simply tell them what the interview will be about, what the interview is for, and how much time it will take.

How to Interview

When interviewing, you have a choice of conducting a traditional, face-to-face interview or an interview using technology. Face-to-face interviews have the strength that you can ask follow-up questions and use non-verbal communication to your advantage. Individuals are able to say much more in a face-to-face interview than in an email, so you will get more information from a face-to-face interview. However, technology provides a host of possibilities when it comes to interviewing people at a distance. You may conduct your interview over the phone, through email, or even virtually through a video-chat program like FaceTime or Skype. You may also use a text or instant messaging program to interview your participant(s), which allows you to ask follow-up questions during the interview and which transcribes the interview for you.

Reducing Bias

One way of eliminating bias in your research is to record your interviews rather than rely on your memory and notes. Before you record any interview, however, be sure that you have permission to record from your participant(s). Recording interviews allows you to directly quote the individual and re-read the interview when you are analyzing and writing. Most computers and cell phones come with recording equipment built in. Taking notes during the interview, however, is a must: this requires you to pay attention, highlight the most important pieces of information, and form new follow-up or clarification questions for your participants.

Once your interview is over, you will need to transcribe your interview to prepare it for analysis. This means creating a written record that is exactly what was said—i.e., typing up your interview(s). If you have conducted an email or chat interview, you will already have a transcription.

Ask about One Thing at a Time

A poorly written question can contain multiple questions, which can confuse participants or lead them to answer only part of the question you are asking. This is called a “double-barreled question.”

Example:

• Double-Barreled Question: What kinds of problems are being faced in the field today and where do you see the search for solutions to these problems going?
• Revised Question 1: What kinds of problems are being faced in the field today?
• Revised Question 2: Where do you see the search for solutions to these problems going?

Avoid Leading Questions

A leading question is one where you prompt the participant to respond in a particular way, which can create bias in the answers given.

Example:

• Leading question: The economy is clearly in a crisis, wouldn’t you agree?
• Revised question: Do you believe the economy is currently in a crisis? Why or why not?

Open Versus Closed Questions

Closed questions, or questions that have yes/no or other limited responses, should be used in surveys. However, avoid these kinds of questions in interviews because they discourage the interviewee from providing in-depth information. In the revised question example above, “Do you believe the economy currently is in a crisis?” could be answered with a simple yes or no, which could keep a participant from talking more about the issue, but the “why or why not?” portion of the question asks the participant to elaborate.

Ask Questions Related to Your Topic

It may seem obvious, but you want to make sure that each question you ask is connected to your research topic/question. Whether you are conducting a survey or an interview, you do not want to waste your or your participants’ time. You want to avoid asking too few questions that do not allow you to gather information and data about your topic, but you also want to avoid asking too many unrelated questions that encourage your participants to abandon the interview before it is finished.

Observations

Observations have lead to some of the most important scientific discoveries in human history. Charles Darwin used observations of the animal and marine life at the Galapagos Islands to help him formulate his theory of evolution that he describes in On the Origin of Species. Today, social scientists, natural scientists, engineers, computer scientists, educational researchers, and many others use observations as a primary research method.

Observations can be conducted on nearly any subject matter, and your research question will determine the kinds of observations you can do. You could observe traffic or parking patterns on campus to get a sense of what improvements could be made. You could observe clouds, plants, or other natural phenomena. If you choose to observe people, you will have several additional considerations, including the manner in which you will both observe them and gain their consent.

Types of Observations

You will likely use unobtrusive observation. In unobtrusive observation, you do not interact with participants but rather simply record their behavior. Although in most circumstances people must volunteer to be participants in research, in some cases it is acceptable not to let participants know you are observing them. In places that people perceive as public, such as a campus food court or a shopping mall, people do not expect privacy, and so it is generally acceptable to observe behavior without participant consent. If it is not practical to get participants’ consent and if your data is anonymous, unobtrusive observations do not violate people’s privacy in public spaces. In places that people perceive as private, which can include a church, home, classroom, or a conversation in a public space, participant consent should be sought.

Participant observation is a method used frequently within ethnographic research in sociology and anthropology. In this kind of observation, a researcher may interact with participants and become part of their community.

Length of Observations

Observations, particularly when conducting ethnographic research, are often conducted over extensive periods of time, in some cases for several years. However, sometimes a research project, such as that completed for a class, does not permit for such extensive data collection. You may only have a few days or even hours to conduct observations, in which case it is important to acknowledge the limited time frame of your data collection in your final report on your research.

Unbiased Observations

The ethical concern of being unbiased is important in recording your observations, which you will do in what is referred to as “field notes.” You need to be aware of the difference between an observation (recording exactly what you perceive through your senses) and an interpretation (making assumptions and judgments about what you see). When you observe, you should focus first on only the events that are directly observable. Consider the following two example field note entries:

1. The student sitting in the dining hall enjoys his greasy, oil-soaked pizza. He is clearly oblivious to the calorie content and damage it may do to his body.
2. The student sits in the dining hall. As he eats his piece of pizza, which drips oil, he says to a friend, “This pizza is good.”

The first entry is biased and demonstrates judgment about the event. First, the observer makes assumptions about the internal state of the student when she writes “enjoys” and “clearly oblivious to the calorie content.” From an observer’s standpoint, there is no way of ascertaining what the student may or may not know about pizza’s nutritional value nor how much the student enjoys the pizza. The second entry provides only the details and facts that are observable.

To avoid bias in your observations, you will want to use an observation protocol. For example, you can use something called a double-entry notebook. This is an approach to recording field notes that encourages you to separate your observations (the facts) from your feelings and judgments about the facts. You simply draw a vertical line down your notebook pages and record your unbiased observations on the left side and your feelings about those events on the right. Some researchers also opt to include a third category in their notebooks for reflection, which enables them to identify why they may have reached the interpretation about the events that they did; for example, a researcher may have experienced something in the past that shaped their perception of an event. You could also use an observation or tallying log, a pre-made table or some other chart that you create before you begin the observation, to help you quickly indicate when certain behaviors or interactions are performed by the people you are observing.