Can an intensive study have a small sample?

Yes; intensity and the number of sampled units are separate.

Can intensive sampling be random?

Yes, if the selection procedure is randomized within the defined population or area.

Does the label specify a formula for uncertainty?

No. That depends on the design, outcome, and analysis.

AI Glossary / Statistics and probability / Advanced

Intensive sampling

Definition

Intensive sampling concentrates observation effort within a limited area or a narrow subject range. In the supplied OECD glossary, it has two meanings: densely distributed sampling points in a particular area, or detailed questioning about a restricted set of topics.

What it means

The term describes concentration and depth, not a complete sampling design. A study can inspect one small area closely or ask participants many detailed questions about one process. Neither choice, by itself, specifies how locations or participants were selected.
Intensity also differs from total sample size. Twenty measurements in a small area can be spatially more intensive than twenty measurements spread across a large region. Similarly, a study of thirty people can examine a topic more deeply than a brief survey of many more people, without becoming more representative of the broader population.

How it works

First identify what is being concentrated: measurement locations, subject coverage, or another explicitly defined aspect of observation. Set the target area or topic and decide what detail is required. Then specify the actual sampling procedure and data-collection method.
For spatial work, document locations, spacing, and the region to which results apply. For detailed questioning, define the participant population, question sequence, response recording, and limits on interview duration. A concentrated design still needs a clear reason for selecting each unit.
Plan how the additional detail will be analyzed. Many questions from one person, or nearby measurements of one local condition, should not automatically be counted as independent evidence about a much wider population.

Three practical examples

A field team places many measurement points within one small inspection area to understand local variation. The dense observations answer a local question rather than automatically representing an entire region.
A user-research team investigates one AI-assisted workflow through detailed interviews. It focuses on specific steps, interruptions, and reasons for decisions instead of covering many unrelated topics superficially.
An operations study concentrates detailed observation on a defined process. It documents the selection of the process and participants so readers can judge how far the findings might transfer.

Think of it this way

Intensive sampling resembles magnifying a small part of a map. More detail becomes visible there, but magnification does not reveal what lies outside it.

Worked example

Consider two invented spatial plans, each using twenty measurement locations. Plan A places them within four hectares. Plan B distributes them across one hundred hectares.
Plan A has twenty divided by four = five locations per hectare. Plan B has twenty divided by one hundred = 0.2 locations per hectare. Plan A's average spatial density is twenty-five times Plan B's, even though both plans have exactly twenty locations.
That calculation describes density, not inferential quality. If Plan A's locations all lie along one accessible edge, the nominal density does not establish good coverage of its four-hectare area. The placement rule and the phenomenon being measured still matter. Neither plan's density alone supplies a margin of error.
For the second meaning, consider a separate interview project with sixty interviewer-hours available. A thirty-minute interview permits 120 interviews if all time is used for interviewing. A two-hour interview permits thirty. These simplified counts exclude recruitment, travel, transcription, and analysis time.
Suppose the short interview covers six topics for about five minutes each. The long interview covers two topics for about sixty minutes each. The longer plan provides twelve times as much interview time per topic per participant: sixty divided by five equals twelve. It also reaches only one-quarter as many participants.
The long plan contains thirty participants and sixty participant-topic discussions. It does not contain sixty independent people, nor does every follow-up question create another sampled participant. Its depth may help explain how a workflow operates while leaving broad prevalence questions unanswered.
Now suppose ten of the thirty participants come from the same team. Their accounts may share practices and experiences. Detailed questioning can reveal that common context, but cannot make the sample equivalent to thirty independently selected organizations.
The spatial plan concentrates observation geographically; the interview plan concentrates it by topic and person. Neither comparison implies that intensive sampling is always preferable; the appropriate balance follows the study objective and selection design.

Illustrative case study

Fictional company Alder Workflow wants to understand why employees interrupt an AI-assisted review process. It chooses detailed interviews about that single workflow rather than a broad questionnaire about every workplace tool.
Researchers record how participants were recruited and distinguish recurring explanations from estimates of how common each explanation is across the company. They use the findings to refine questions for a later, broader study. Interview depth helps uncover mechanisms and vocabulary; it does not establish organization-wide frequencies. This case is fictional.

Common misconceptions and mistakes

Intensive does not mean statistically representative. Selection, coverage, and nonresponse must be considered separately.
More observations in one area do not necessarily resolve uncertainty about distant areas. Local detail and broad coverage are different properties.
A lengthy questionnaire is not automatically a better instrument. Additional questions can be repetitive, confusing, or burdensome if they do not serve the study objective.
The term is not synonymous with purposive, random, stratified, or cluster sampling. Those labels describe other features of a design and may coexist with intensive collection.
Frequent repeated measurements are sometimes described as intensive in particular fields. The present entry preserves the supplied glossary's spatial-density and topic-depth meanings; specialized temporal usage should be defined explicitly.

Business use

Use an intensive approach when a decision needs detailed understanding of a defined setting or process. Specify the boundary of that setting, explain recruitment or location selection, and budget for analysis as well as collection. Present detailed findings with their scope. If the decision also requires population proportions, design an appropriate broader sampling component rather than treating depth as a substitute for coverage.

Check your understanding

Which spatial plan has higher location density?

Answer: Plan A, with five locations per hectare.

How many people can receive two-hour interviews in sixty interviewing hours?

Answer: Thirty, under the simplified assumptions.

Do sixty participant-topic discussions equal sixty independent participants?

Answer: No; they come from thirty people.

Related terms

Extensive sampling
Sampling design
spatial sampling
sampling unit
coverage
participant selection
questionnaire design
generalizability

Learn next

Compare depth with breadth in study design. Study probability sampling, spatial coverage, respondent burden, and methods for analyzing repeated or clustered observations.

Sources

OECD Glossary of Statistical Terms, 2008, Intensive sampling, citing Marriott's Dictionary of Statistical Terms: https://www.oecd.org/content/dam/oecd/en/publications/reports/2008/09/oecd-glossary-of-statistical-terms_g1gh9ad7/9789264055087-en.pdf . Spatial plans, interview budgets, and company scenario are original illustrations.

Previous: Intensional Component Next: Intention-to-treat