Isolation Index
Also known as: P* isolation index, interaction index, exposure index, Bell isolation index
The isolation index measures the exposure dimension of segregation: the extent to which members of a minority group are exposed only to one another rather than to members of other groups. It answers the question 'what is the own-group share of the typical neighbor (or classmate, or coworker) that a member of the focal group encounters?' Unlike evenness measures, it depends on the relative size of the group as well as its spatial distribution.
Key highlights
- Directly captures lived exposure — the group composition members actually encounter — which evenness measures miss.
- Interaction and isolation forms together describe both within- and cross-group contact, summing to one in the two-group case.
- Simple to compute from standard cross-tabulated unit counts.
- The standard measure of the exposure dimension, complementing dissimilarity's evenness.
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 the isolation (and interaction) index when the substantive concern is exposure — the actual chance of within-group versus cross-group contact in neighborhoods, schools, or workplaces — rather than mere evenness. It is the standard measure of the exposure dimension and is appropriate when group size itself matters for the experience of segregation. It is less appropriate when you want a size-independent measure of evenness (use dissimilarity) or a multigroup decomposition (use the entropy index), and the raw index can be misleading when comparing groups of very different sizes unless the normalized form is used. Like other aspatial indices, it ignores the arrangement of units in space.
Strengths & limitations
- Directly captures lived exposure — the group composition members actually encounter — which evenness measures miss.
- Interaction and isolation forms together describe both within- and cross-group contact, summing to one in the two-group case.
- Simple to compute from standard cross-tabulated unit counts.
- The standard measure of the exposure dimension, complementing dissimilarity's evenness.
- Depends on the group's overall size, so a small group can show low isolation even when highly clustered, unless normalized.
- Aspatial: ignores where units lie relative to one another.
- The basic form handles exposure between two groups; multigroup exposure needs generalized P* matrices.
- Sensitive to unit definition and the modifiable areal unit problem like other segregation indices.
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 does the isolation index differ from the index of dissimilarity?
Dissimilarity measures evenness — how unequally two groups are spread across units — and is independent of group size. Isolation measures exposure — the own-group composition a typical member encounters — and depends on group size as well as distribution. A small minority can be highly uneven (high dissimilarity) yet have low isolation simply because there are too few of its members to fill any unit. The two capture distinct dimensions of segregation and should be reported together.
What is the interaction index?
The interaction index is the exposure counterpart that measures the average proportion of a focal group member's unit made up of another group — that is, potential cross-group contact. In the two-group case isolation and interaction sum to one. Interaction indices are useful when the question is about intergroup contact rather than within-group insulation.
Why does isolation depend on group size?
Because the index averages the actual own-group share in members' units, a group that is only a few percent of the total population cannot, even under perfect evenness, make up a large share of any unit. Its isolation is therefore mechanically low. The normalized eta-squared form removes this size floor so that the purely distributional component can be compared across groups and places.
Sources
- 1.Bell, W. (1954). A probability model for the measurement of ecological segregation. Social Forces, 32(4), 357–364.
- 2.Massey, D. S., & Denton, N. A. (1988). The dimensions of residential segregation. Social Forces, 67(2), 281–315.
You have read it. What now?
Cite this page
ScholarGate. (2026, June 22). Isolation Index. ScholarGate. https://scholargate.app/sociology/isolation-index