Within-and-Between Analysis
Also known as: WABA, Within and Between Entities Analysis, Dansereau WABA, Levels-of-Analysis Analysis
Within-and-Between Analysis (WABA) is a methodology for determining the level of analysis at which a relationship between variables actually operates, developed by Fred Dansereau, Joseph Alutto, and Francis Yammarino in their 1984 book on the varient approach to theory testing. The central question it answers is whether an observed correlation reflects a group-level phenomenon (differences between work units), an individual-level phenomenon (differences among individuals within units), both, or neither. WABA decomposes the variance of each variable, and the covariance between variables, into between-entity and within-entity components, then applies statistical and practical tests to draw a levels inference. Yammarino and Markham's 1992 application showed how WABA can overturn casual assumptions, demonstrating that phenomena presumed to be group-based may in fact be individual-based. Klein, Dansereau, and Hall's 1994 review situated WABA within a broader argument that levels of analysis must be specified in theory, measurement, and analysis alike. WABA forces researchers to test, rather than assume, the level at which their constructs live.
Key highlights
- Forces explicit specification and empirical testing of the level of analysis rather than assuming it.
- Decomposes any observed correlation into interpretable between-group and within-group contributions.
- Combines statistical tests with practical-significance criteria, guarding against significance driven purely by sample size.
- Provides a clear classification (wholes, parts, equivocal, inexplicable) that maps directly onto theoretical levels claims.
Intuition
This section is available to Pro members. Upgrade to Pro
How it works
This section is available to Pro members. Upgrade to Pro
When to use it
Use within-and-between analysis when individuals are nested in groups and you need to establish whether a relationship operates at the group level, the individual level, or both before interpreting it or building theory on it. It is appropriate for questions about climate, leadership dyads, team processes, and any construct whose level is contested, and it pairs naturally with theory that makes explicit levels claims. It is less suited to designs without a meaningful nesting structure, to questions where the level is already firmly established, or to situations requiring estimation of random effects and cross-level interactions, where multilevel (hierarchical linear) modeling is the more flexible tool. WABA is best seen as a disciplined diagnostic for levels inference rather than a general-purpose modeling framework.
Strengths & limitations
- Forces explicit specification and empirical testing of the level of analysis rather than assuming it.
- Decomposes any observed correlation into interpretable between-group and within-group contributions.
- Combines statistical tests with practical-significance criteria, guarding against significance driven purely by sample size.
- Provides a clear classification (wholes, parts, equivocal, inexplicable) that maps directly onto theoretical levels claims.
- It is primarily a diagnostic for levels inference and does not estimate random effects or cross-level interactions the way multilevel modeling does.
- Results are sensitive to the choice of entity, and a poorly justified grouping yields an uninterpretable decomposition.
- The practical-significance benchmarks of the E-test are conventions and have been debated.
- It handles two-level structures cleanly but becomes cumbersome for designs with more than two nested levels.
Common pitfalls
This section is available to Pro members. Upgrade to Pro
Applications
This section is available to Pro members. Upgrade to Pro
Frequently asked
How is WABA different from hierarchical linear modeling?
Both deal with nested data, but they answer different questions. WABA is a diagnostic that decomposes variance and covariance into between-group and within-group parts and classifies the level at which a relationship operates, combining statistical and practical-significance tests. Hierarchical linear (multilevel) modeling is an estimation framework that fits random intercepts and slopes and estimates cross-level interactions. In practice, researchers often use WABA-style reasoning to establish the level of a construct and multilevel modeling to estimate the relationships once the level is understood; they are complementary rather than competing.
What do the labels wholes, parts, and equivocal mean?
They are WABA's levels classifications. 'Wholes' means the relationship is a group-level phenomenon, driven by differences between groups, so individuals within a group are essentially interchangeable on that variable. 'Parts' means the relationship is an individual-level phenomenon, driven by differences among individuals within their groups, with the grouping irrelevant. 'Equivocal' means both the between-group and within-group components contribute meaningfully, so the relationship operates at both levels. There is also an 'inexplicable' label for when neither component shows meaningful covariation, signaling that WABA finds no level at which the relationship holds.
Why combine statistical tests with practical-significance criteria?
Because statistical significance is sensitive to sample size and can flag trivial differences as meaningful in large datasets. Yammarino and Markham stress pairing the F-test with the E-test and practical benchmarks so that a between-group component is accepted only if it is both unlikely to be chance and large enough to matter substantively. This dual standard is central to WABA's philosophy: the goal is a defensible levels inference about how the world actually works, not merely a p-value, so both kinds of evidence must point the same way before a level is claimed.
Sources
- 1.Dansereau, F., Alutto, J. A., & Yammarino, F. J. (1984). Theory Testing in Organizational Behavior: The Varient Approach. Prentice-Hall.ISBN 9780133595079
- 2.Yammarino, F. J., & Markham, S. E. (1992). On the application of within and between analysis: Are absence and affect really group-based phenomena? Journal of Applied Psychology, 77(2), 168-176.
- 3.Klein, K. J., Dansereau, F., & Hall, R. J. (1994). Levels issues in theory development, data collection, and analysis. Academy of Management Review, 19(2), 195-229.
You have read it. What now?
Cite this page
ScholarGate. (2026, June 23). Within-and-Between Analysis. ScholarGate. https://scholargate.app/organizational-behavior/within-and-between-analysis