Defining the key unit of analysis is an integral yet often overlooked component of qualitative research design. As the fundamental entity being studied and drawing conclusions about, the unit of analysis profoundly shapes research outcomes and the validity of findings.

This article will clarify exactly what constitutes a unit of analysis, compare units of observation, outline different levels of analysis, and provide guidance on appropriately selecting units of analysis for meaningful research.

What is a Unit of Analysis?

The unit of analysis refers to the primary focus of data collection and analysis in a study. It is the ‘what’ or ‘who’ that forms the basis for research hypotheses, data gathering, analytical procedures, and drawing conclusions. Units of analysis are, therefore, central building blocks in qualitative research.

For example, a study exploring the effects of mindfulness practices on employee stress would have the individual employee as the unit of analysis. Data will be collected about employees’ stress levels, perceptions of workplace demands, and frequency of mindfulness techniques. The study would then assess how these factors relate to one another at the employee level.

Meanwhile, a study on increasing gender diversity in Silicon Valley technology firms would have the organization as the unit of analysis. Data about hiring practices, promotion rates, and representation figures would be gathered for companies, and conclusions drawn about relationships between diversity policies and women’s advancement, specifically within firms.

Units of Observation vs. Units of Analysis

There is often confusion between different units of analysis and units of observation in qualitative studies. While closely related, they refer to slightly different aspects of the research:

  • Unit of observation – The source of the data collected in the study, for example, individuals, documents, settings
  • Unit of analysis – The entity that data is analyzed and conclusions are drawn about, for example, individuals, groups, organizations

For instance, a study exploring children’s attitudes towards climate change may gather data by interviewing parents (unit of observation), but the analysis focuses on understanding perspectives amongst the children themselves (unit of analysis). Getting clarity on the sometimes subtle differences between observation and analysis units is important for maintaining consistency when collecting, assessing, and reporting on research data.

Types of Level of Analysis

Units of analysis in qualitative research typically fall within five key levels:

Individual Level 

Studies with individuals as the unit of analysis collect and analyze data focused on individual people, e.g., personal attitudes, lifestyle factors, and decision-making drivers. Data can be gathered through interviews, surveys, observation, or review of personal records or accounts held by an individual. Examples of research questions explored at this level include investigating employee retention issues by understanding the perspectives and experiences of individual workers.

Aggregates Level

This level investigates collections of individuals without assessing the views or characteristics of specific persons within those groups. Data provides aggregated summaries only about particular groups. Sub-categories within this level include:

Groups – Teams, families, friendships circles, etc. An example research question would be examining collaboration practices amongst customer service teams.

Organizations – Companies, government agencies, universities, etc. For instance, evaluating procurement policies amongst hospitals.

Social Level

The societal level explores phenomena at a wider cultural, institutional, or public level beyond specific organizations or grouped individuals. Sub-categories here include:

Social Artifacts Level – Products, frameworks, systems, policies, processes, etc. An example would be reviewing the social impacts of renewable energy programs and infrastructure.

Social Interaction Level – Communications, relationships, exchanges, etc. For instance, assessing evolving intercultural attitudes and engagement between immigrant and local communities over time.

Importance of Selecting the Correct Unit of Analysis in Research

Choosing an appropriate unit of analysis is crucial for producing valid, reliable, and high-quality research outcomes. Reasons why getting the unit of content analysis right matters include:

  • Ensures alignment with research aims and questions. The unit of analysis channels attention toward collecting data that helps social scientists answer proposed research questions.
  • Enables drawing meaningful conclusions. Conclusions can only be drawn about the entity defined as the unit of analysis (not about wider or different populations).
  • Reveals true relationships between factors. Defining the unit of analysis clarifies the scope of measured relationships between different variables or themes.
  • Allows for repeatability and comparison. Consistently defining units of analysis enables different studies and datasets to be accurately reproduced and compared over time.

Tightly defining units of analysis strengthens overall research rigor and the validity of conclusions.

Examples of a Unit of Analysis

To illustrate the concept further, here are some examples of studies with well-defined units of analysis:

  • Research question: How effective are smartphone apps at motivating increased physical activity among university students? Unit of analysis: Individual students
  • Research question: Do companies with female CEOs have better environmental sustainability performance? Unit of analysis: Organizations (companies)
  • Research question: How financially literate are fourth-grade children from low socioeconomic status communities? Unit of analysis: Grade four student cohorts

As shown above, honing in on the specific entity that findings apply to helps frame research around well-constructed and suitably targeted questions. This then sets up the study’s analytical procedures and conclusions to be robust and meaningful.

Factors to Consider When Choosing a Unit of Analysis

When selecting appropriate units of analysis, four key factors should guide decision-making:

Research Questions and Hypotheses

The first reference point is whether the proposed unit of analysis would help answer the study’s core research questions and hypotheses. Units of analysis provide an orienting framework and filter for collecting data that helps prove or disprove posed hypotheses.

Data Availability and Quality

The availability of reliable and relevant data on potential units of analysis is a pragmatic consideration. Can quality data realistically be sourced given constraints like costs, access barriers, or collection complexity? The richness and depth of available data impact analysis quality.

Feasibility and Practicality

Research capacity in terms of timeline, budget, resourcing, and technical capabilities also determines what units of analysis are feasible. For example, multi-year analyses of nationwide public policy changes may not be practical for undergraduate research projects. Clarifying realistic parameters helps match units of analysis to available research resources and skills.

Theoretical Frameworks and Research Design

The overall research plan and adopted conceptual models, theories or frameworks guide choosing coherent units of analysis. For instance, studies assuming an interpretive perspective focussing on constructivist meaning-making would typically adopt smaller-scale units of analysis like individuals or groups rather than entire organizations or institutions.

Carefully weighing up these key elements facilitates the selection of well-matched, relevant, and workable units of analysis for a given study context.

Common Mistakes When Choosing a Unit of Analysis

While defining units of analysis is a fundamental research design task, confusion and ambiguity often arise. Two common pitfalls are:


Assuming that knowledge about individuals inevitably informs understanding of grouped behaviors or wider social contexts oversimplifies complex dynamics. For example, assessing personal perceptions does not necessarily reveal full organizational viewpoints – a form of reductionism flattening critical detail. Hence, clearly distinguishing units of analysis avoids inaccurate generalization.

Ecological Fallacy

The opposite assumption – that learning about collectives reveals information about individuals within those groups – also misrepresents relationships. This ‘ecological fallacy’ masks diversity, inferring there is uniformity between layers of analysis when differences usually exist. Again, tightly defining units of analysis, rather than presuming parallels, iteratively improves analysis specificity.


Identifying suitable units of analysis in social sciences is a foundational yet commonly underspecified component within qualitative research design. Units of analysis provide an orienting basis for collecting targeted data that helps answer specified research questions.

Defining the “who” or “what” being studied also enables valid conclusions tightly linked to analyzed entities. Carefully determining appropriate units of analysis early on leverages opportunities for expansive, multi-dimensional, and conclusive findings.