03 · What You Need to Know
How Theoretical Sampling Connects Data Collection with Theory Development
What Exactly Is Theoretical Sampling?
Theoretical sampling is an iterative approach to selecting additional data sources according to the analytical needs of an emerging theory.
In grounded theory, researchers collect and analyze data concurrently. Early analysis generates concepts and categories, which subsequently guide decisions about where to look for further evidence.
For example, initial interviews about AI-supported assessment may suggest that instructors distinguish between technical confidence and institutional permission when deciding whether to use generative AI.
The researcher may then seek participants whose experiences help clarify whether these are separate processes, how they interact, and under which conditions they become consequential.
The selection is theoretical because it follows a developing analytical explanation rather than a fixed demographic allocation.
Theoretical sampling is therefore a process of pursuing conceptual questions through additional evidence.
Why Is Theoretical Sampling Important in Grounded Theory?
Grounded theory seeks to develop theoretical explanations grounded in systematically collected and analyzed data.
Theoretical sampling is important because emerging categories often require further examination before their properties and relationships can be adequately explained.
Glaser and Strauss established theoretical sampling as a central component of the constant comparative approach. Later grounded theory traditions, including those associated with Strauss and Corbin and with Charmaz, interpret the researcher's analytical role and theoretical development in somewhat different ways.
Despite these differences, theoretical sampling generally involves using developing analysis to determine subsequent data collection.
A study claiming to follow grounded theory while collecting all interviews according to an entirely fixed plan and conducting analysis only afterward may need to explain how its design accommodates the iterative logic of the selected grounded theory tradition.
The requirement should be evaluated within the specific methodological approach rather than imposed identically on every qualitative study.
How Is Theoretical Sampling Different from Purposive Sampling?
Purposive sampling is a broad approach in which researchers deliberately select cases because they are relevant to the research question.
Theoretical sampling is more specifically directed by concepts and relationships emerging during analysis.
Initial purposive sampling
Selects cases according to the initial research question, relevant experiences, or contextual characteristics.
Theoretical sampling
Selects subsequent data sources according to what emerging theoretical categories require for further development.
For example, a researcher may initially recruit teachers with direct experience implementing AI-generated feedback.
After analysis identifies institutional trust as a potentially important category, the researcher may seek instructors who experienced policy reversals or conflicting institutional guidance.
The later selection is not simply another purposive recruitment decision. It is directed toward examining the properties and relationships of an emerging category.
Not every purposive adjustment is theoretical sampling, particularly when no developing conceptual explanation guides the decision.
When Does Theoretical Sampling Begin?
Theoretical sampling begins when analysis provides a substantive reason to seek additional evidence.
Researchers may begin with an initial purposive sample selected according to the broad phenomenon of interest.
Early interviews, observations, documents, or other data are analyzed to identify provisional concepts and questions.
These questions guide subsequent selection.
For example, initial interviews may suggest that teachers distinguish between using AI privately for preparation and using it visibly in student assessment.
The researcher might then seek cases involving public disclosure, formal assessment approval, or institutional monitoring to clarify that distinction.
Theoretical sampling should not be postponed until all originally planned data have been collected if the methodological design requires concurrent analysis and recruitment.
What Should Guide the Selection of the Next Case?
The next case should be selected according to a specific analytical need.
Researchers might seek evidence that clarifies a category's properties, identifies conditions under which a process changes, examines relationships between categories, or challenges an emerging interpretation.
For example, a developing category called perceived professional risk may require evidence from instructors working under different institutional disclosure requirements.
The researcher should ask what additional case would help distinguish professional risk from related concepts such as uncertainty, distrust, or perceived workload.
Such decisions are strengthened by analytical memos documenting the category, the unresolved question, and the reason for selecting a particular data source.
Theoretical sampling is not a search for participants who will confirm the emerging theory. Disconfirming evidence may be especially valuable.
Can Theoretical Sampling Involve Documents or Events Rather Than Participants?
Yes. Theoretical sampling concerns the selection of evidence, not necessarily the recruitment of additional interviewees.
Researchers may select policy documents, organizational settings, observed interactions, incidents, or other materials according to the needs of emerging categories.
For example, interviews may suggest that institutional AI policy revisions influence teachers' assessment decisions.
The researcher might examine successive versions of institutional policy documents to clarify the sequence and nature of those revisions.
Alternatively, observations of assessment meetings may provide evidence about how policy uncertainty is negotiated collectively.
The appropriate data source depends on what the emerging theoretical question requires and what the study's methodological and ethical arrangements permit.
How Does Constant Comparative Analysis Inform Theoretical Sampling?
Constant comparative analysis involves comparing incidents, accounts, and developing categories to identify similarities, differences, properties, and relationships.
These comparisons may reveal conceptual gaps that require further evidence.
Suppose two instructors report similar institutional AI policies but make different assessment decisions.
Comparison may suggest that formal policy alone does not explain their behavior.
The researcher may then investigate differences in perceived autonomy, professional accountability, or previous experiences with institutional enforcement.
Theoretical sampling follows from this analytical problem by selecting cases that can help clarify the proposed explanation.
Data collection and analysis consequently inform one another rather than proceeding as completely separate stages.
What Role Do Analytical Memos Play?
Analytical memos provide a record of developing concepts, questions, comparisons, and decisions.
They can help researchers distinguish genuine theoretical sampling from recruitment changes driven primarily by convenience.
A useful memo might identify the emerging category, explain what remains unclear, propose a type of case or data source, and specify what evidence would challenge the current interpretation.
For example, a memo might state that the category institutional trust appears related to teachers' willingness to disclose AI use, but the relationship remains unclear in settings with highly restrictive policies.
The next sampling decision could therefore seek instructors from such settings.
Memos should remain analytical records rather than retrospective justifications written only after recruitment has ended.
How Is Theoretical Sampling Different from Maximum Variation Sampling?
Maximum variation sampling deliberately selects cases differing across predefined characteristics considered relevant to the research question.
Theoretical sampling selects additional evidence according to the developing analysis.
For example, maximum variation sampling might initially include instructors from public and private universities, different disciplines, and different AI policy environments.
Theoretical sampling might later focus specifically on instructors who experienced a policy reversal because that event appears important to an emerging explanatory category.
Theoretical sampling can involve contrasting cases, but contrast alone does not make a decision theoretical.
The distinction concerns whether selection is guided by the emerging theory.
Does Theoretical Sampling Require Seeking Negative Cases?
Negative or disconfirming cases may be valuable when an emerging category or relationship appears too simple.
Suppose the researcher proposes that greater institutional trust encourages instructors to use AI-generated feedback.
An instructor who reports high trust but avoids AI feedback may challenge that relationship.
Investigating the case could reveal additional conditions, such as disciplinary assessment requirements or concerns about feedback reliability.
However, theoretical sampling is broader than negative-case selection. It may also seek cases that elaborate category properties or clarify relationships not yet sufficiently understood.
Negative and deviant case sampling can contribute to theoretical development when it responds to a specific analytical need.
How Do You Know When Theoretical Sampling Is Sufficient?
In grounded theory, theoretical saturation is commonly used to describe the point at which further data collection is no longer contributing substantively to the development of relevant theoretical categories.
However, the meaning and operationalization of saturation differ across grounded theory traditions.
Theoretical saturation should not be reduced to the absence of new interview topics or codes.
A category may appear repeatedly while its conditions, properties, consequences, or relationships remain insufficiently understood.
Researchers should evaluate whether the theoretical account is adequately developed for the study's purpose and whether important conceptual questions remain unresolved.
Practical constraints may also require recruitment to stop before full theoretical development is achieved. Such limitations should be reported rather than relabeled as saturation.
Is Theoretical Saturation the Same as Code Saturation?
No. Code saturation generally concerns the identification of recurring codes or topics, whereas theoretical saturation concerns the development of categories and their relationships within a theoretical account.
Hennink, Kaiser, and Marconi distinguish code saturation from meaning saturation, demonstrating why identifying recurring topics does not necessarily establish sufficient depth of understanding.
Theoretical saturation adds the grounded theory concern with conceptual development and integration.
Researchers should therefore avoid using the statement that no new codes emerged as the sole justification for stopping theoretical sampling.
The stopping rationale should identify what has been sufficiently developed and which analytical uncertainties remain.
How Many Participants Does Theoretical Sampling Require?
There is no universal number.
Theoretical sampling is directed by emerging analytical needs rather than a predetermined participant count.
Some categories may require examination across several contrasting cases, while others may become sufficiently developed with fewer additional sources.
The complexity of the phenomenon, quality of data, analytical approach, and scope of the developing theory all matter.
Researchers may establish initial recruitment plans for feasibility and ethics purposes, but should avoid treating a fixed number as proof of theoretical saturation.
When resource limitations restrict further sampling, the resulting theoretical limitations should be acknowledged.
Is Theoretical Sampling Necessary in Every Qualitative Study?
No. Theoretical sampling is especially associated with grounded theory and is not a universal requirement across qualitative methodologies.
A phenomenological study may use criterion-based or homogeneous sampling to investigate a shared lived experience.
A qualitative case study may select cases according to their relevance to a bounded institutional phenomenon.
These approaches may permit iterative recruitment without requiring the particular logic of grounded theory theoretical sampling.
Researchers should avoid using theoretical sampling as a prestigious label for any change in recruitment.
The term should be reserved for selection decisions genuinely directed toward developing emerging theoretical categories.
How Should Theoretical Sampling Be Reported?
The methodology should explain the initial sampling strategy, the concurrent analysis procedures, and how emerging categories informed subsequent selection.
Researchers should provide examples of analytical questions that prompted additional recruitment or data collection.
They should also describe the types of cases sought, the evidence obtained, and how that evidence contributed to category development.
Where theoretical saturation is claimed, the explanation should address the development of categories and relationships rather than rely solely on a participant count.
Reporting should be consistent with the particular grounded theory tradition and transparent about practical constraints.
