Surveys, interviews, and ethnography: choose evidence for the question

MCAT content library / Research & data / Surveys, interviews, and ethnography: choose evidence for the question

Choose a method by the information a question needs, then keep the conclusion within what the resulting observations can show.

About 14 minutes, including the worked examples and checks.

Useful foundations

• Psychological Measurement: What Does the Score Mean?

On this page

• Concept: start with the answer you need
• Pattern: distinguish a report from an observation
• Application: a campus resource that students avoid
• Rule: choose the next method to resolve the remaining uncertainty
• Inspect the instrument and the observer
• Keep inference proportional to the evidence
• Check your reasoning
• Continue learning
• Sources

Concept: start with the answer you need

A method is useful when its observations can answer the research question. “How many students used the tutoring center?” requires a count with a defined population and time interval. “What does asking for tutoring mean to students?” asks about interpretations. “What happens when a student first approaches the center?” asks about a process in a setting. These questions may belong to the same project, but they do not call for identical evidence.

A structured survey asks comparable questions across participants. An interview can probe an account, clarify language, and follow an unexpected explanation. Ethnographic work uses sustained engagement with a social setting to understand practices and meanings in context, often combining observation, conversation, and records. These labels describe approaches; none guarantees representativeness, truthfulness, or causal identification.¹

Pattern: distinguish a report from an observation

If a questionnaire asks whether someone felt unwelcome, the answer records that person’s report under the survey conditions. It does not directly observe every interaction that produced the feeling. If an observer records that staff did not greet three arriving students, the notes describe visible behavior in those encounters. They do not, by themselves, establish the students’ private interpretation or the staff’s intention. Combining accounts and observations can make a proposed explanation more specific without making the two kinds of evidence interchangeable.

Likewise, “qualitative” and “quantitative” describe the form and use of information more than a rigid division between methods. An interview can collect a numerical waiting time and an extended explanation. A survey can include an open-ended response. Ethnographic field notes can contain event counts. Ask what was actually collected and how it was analyzed.¹

Application: a campus resource that students avoid

All scenarios and data on this page are original teaching examples. A college wants to understand low use of a free tutoring center. Three investigators produce the following evidence.

Approach	Observed result	Immediate inference limit
Survey of 300 randomly invited respondents	45% report that opening hours conflict with work.	Respondents may differ from invited nonrespondents.
Interviews with 18 purposively selected nonusers	Several describe the center as intended only for students already failing.	The frequency of this view in all students is not established.
Six weeks of observation during varied opening hours	New arrivals often pause at an unlabeled desk before leaving.	The observations do not establish why each person left.

The survey provides a standardized estimate of reported scheduling difficulty among its respondents. To extend 45% to the target population, examine the frame, selection process, response pattern, question wording, and weighting. Random invitations alone do not make nonresponse harmless. The question also measures reported conflict, not the causal effect of opening later.

The interviews reveal a possible barrier the initial questionnaire may have missed: the meaning attached to using tutoring. Selecting participants for a range of experiences can be useful when the goal is to discover and develop explanations. It does not turn 18 accounts into a probability estimate. “Several interviewees described stigma” fits the evidence. “Most students avoid tutoring because of stigma” requires additional information.

The observation suggests a concrete process worth investigating: uncertainty at the entrance. Follow-up questions can ask departing students what they expected and what prevented them from continuing. A better sign might help, but observing hesitation is not a trial of signage. Work schedules, wrong-room arrivals, or a canceled appointment could produce the same visible behavior.

Rule: choose the next method to resolve the remaining uncertainty

If the next question is how widespread the newly identified interpretation is, develop clear survey items from the interview findings and test them in an appropriate sample. Check that students understand the wording as intended. If the question is how the interpretation develops during interactions, repeated observation and interviews may be more informative than another frequency count. If the question is whether a new entrance procedure increases use, a comparative design that varies that procedure is needed.

A mixed-method sequence can therefore move from discovery to measurement to testing. The benefit comes from each method addressing a gap in the preceding evidence. Merely collecting more types of data does not automatically strengthen a claim. If all three approaches exclude working students, their agreement may reflect a shared sampling limitation.

Inspect the instrument and the observer

Consider the survey item “How much did the center’s inconvenient hours prevent you from seeking the help you needed?” It assumes that the hours were inconvenient, that help was needed, and that attendance was prevented. An answer cannot cleanly distinguish those assumptions. Separate the ideas: whether help was sought, which hours were available, and whether the schedule conflicted with other obligations. Clearer wording improves the match between the question and the intended construct.

Interviewers can also steer accounts by supplying explanations too early. “Why were you embarrassed?” presupposes embarrassment. “What were you thinking when you considered attending?” leaves more room for an unexpected response. Observers should specify what counts as a pause or a departure, document the setting and their role, and distinguish field observations from later interpretations. Agreement between coders helps assess consistency, but shared agreement alone does not prove that the coding scheme captures the intended construct.²

Keep inference proportional to the evidence

Two contrasting accounts are not necessarily a reason to discard one. A student may value tutoring in an interview yet avoid entering when peers are nearby. The mismatch may reveal context-dependent behavior, social desirability in one setting, or an incomplete question. Investigate the conditions under which each observation was obtained.

For a method-selection question, name the missing information first. A representative frequency estimate, a detailed account of meaning, an observed social process, and a causal intervention effect are different targets. Choose the method that collects the relevant evidence, then carry its limits into the conclusion. That sequence is more reliable than ranking surveys, interviews, and ethnography from strongest to weakest without considering the question.

Check your reasoning

Choose an answer before opening its feedback. Each option explains the reasoning behind that choice.

Continue learning

Related lessons

• Sampling and generalizability: trace who enters the estimate
• Research ethics: match the safeguard to the problem

Browse research & data · MCAT content library

MCAT preparation and roadmap · Aim High, Score Higher

Sources

1. OpenStax Introduction to Sociology 3e — Research Methods

2. OpenStax Psychology 2e — Approaches to Research

3. AAMC — What is on the MCAT exam?