Intersecting Diversity in Combinatorics & Social Metrics

• Intersecting Diversity is a dual concept that measures the deviation of intersecting families from star configurations in combinatorics and the probability of trait differences in social groups.
• It employs advanced techniques such as shifting, junta methods, and cross-intersecting results to establish benchmarks and stability in extremal settings.
• Extensions to weighted, higher-order, and symmetric-difference variants expand its applications from subspace arrangements to networked social and demographic analyses.

In extremal combinatorics, intersecting diversity denotes a family of parameters that quantify how far an intersecting family is from a star, that is, from the most concentrated Erdős–Ko–Rado-type extremal configuration. In a distinct quantitative social-science usage, intersecting diversity is a normalized probability that two sampled group members differ in at least one trait when identities are treated jointly across several axes. Both usages replace one-dimensional size counts by measures of spread, concentration, or aggregate identity composition, but they do so in different mathematical settings and with different extremal or probabilistic objectives (Frankl et al., 2018, Hoogstra et al., 11 Aug 2025).

1. Core combinatorial meaning

Let $[n]=\{1,\dots,n\}$, and let $\mathcal F\subseteq \binom{[n]}{k}$. The family $\mathcal F$ is intersecting if

$$F\cap F'\neq \emptyset \qquad \text{for all }F,F'\in \mathcal F.$$

A full star is

$\mathcal S_x=\{F\in \tbinom{[n]}{k}: x\in F\},$

and any subfamily of a full star is a star. The degree of an element $i\in[n]$ is the number of members of $\mathcal F$ containing $i$, and the maximum degree is

$\Delta(\mathcal F)=\max_{i\in[n]} d_i.$

The standard diversity is

$$\gamma(\mathcal F)=|\mathcal F|-\Delta(\mathcal F),$$

equivalently,

$\mathcal F\subseteq \binom{[n]}{k}$0

Thus diversity is exactly the number of sets avoiding a most popular element; it is $\mathcal F\subseteq \binom{[n]}{k}$1 precisely for stars, so it measures how much of the family lies outside a largest star (Kupavskii, 2017, Frankl et al., 2018).

This parameter is natural because the Erdős–Ko–Rado theorem identifies stars as the largest intersecting $\mathcal F\subseteq \binom{[n]}{k}$2-uniform families. Diversity asks a different extremal question: not how large an intersecting family can be, but how large its non-star residue can be. In shifted families, the degrees satisfy $\mathcal F\subseteq \binom{[n]}{k}$3, and then $\mathcal F\subseteq \binom{[n]}{k}$4 is simply the number of sets avoiding $\mathcal F\subseteq \binom{[n]}{k}$5; this makes diversity a stability parameter as well as an extremal one (Frankl et al., 2018).

2. Ordinary diversity and its extremal constructions

A central benchmark is Frankl’s conjectured bound

$\mathcal F\subseteq \binom{[n]}{k}$6

motivated by the classical “two out of three” family

$\mathcal F\subseteq \binom{[n]}{k}$7

for which

$\mathcal F\subseteq \binom{[n]}{k}$8

The same binomial scale also appears in the “pure triangle” family

$\mathcal F\subseteq \binom{[n]}{k}$9

which has ordinary diversity exactly $\mathcal F$0 (Kupavskii, 2017, Frankl et al., 2023).

The asymptotic large-$\mathcal F$1 theory was developed in stages. A 2017 result proved that there exists an absolute constant $\mathcal F$2 such that for $\mathcal F$3, every intersecting $\mathcal F$4 satisfies

$\mathcal F$5

and if equality holds then $\mathcal F$6 is a subfamily of an isomorphic copy of $\mathcal F$7 (Kupavskii, 2017). A later improvement showed that if $\mathcal F$8, then every intersecting family satisfies

$\mathcal F$9

improving the previous best threshold $$F\cap F'\neq \emptyset \qquad \text{for all }F,F'\in \mathcal F.$$0; the proof proceeds through strong lower bounds on $$F\cap F'\neq \emptyset \qquad \text{for all }F,F'\in \mathcal F.$$1 for large intersecting families (Frankl et al., 2023).

The conjectured threshold $$F\cap F'\neq \emptyset \qquad \text{for all }F,F'\in \mathcal F.$$2, however, is not valid in full generality. For sufficiently large $$F\cap F'\neq \emptyset \qquad \text{for all }F,F'\in \mathcal F.$$3 and

$$F\cap F'\neq \emptyset \qquad \text{for all }F,F'\in \mathcal F.$$4

there exists an intersecting family $$F\cap F'\neq \emptyset \qquad \text{for all }F,F'\in \mathcal F.$$5 such that

$$F\cap F'\neq \emptyset \qquad \text{for all }F,F'\in \mathcal F.$$6

which disproves Frankl’s $$F\cap F'\neq \emptyset \qquad \text{for all }F,F'\in \mathcal F.$$7 threshold and also a stronger conjecture of Kupavskii in the $$F\cap F'\neq \emptyset \qquad \text{for all }F,F'\in \mathcal F.$$8 case (Huang, 2018). In the nonuniform setting $$F\cap F'\neq \emptyset \qquad \text{for all }F,F'\in \mathcal F.$$9, Huang’s conjecture was likewise disproved by explicit intersecting families $\mathcal S_x=\{F\in \tbinom{[n]}{k}: x\in F\},$0 and $\mathcal S_x=\{F\in \tbinom{[n]}{k}: x\in F\},$1 whose diversity exceeds

$\mathcal S_x=\{F\in \tbinom{[n]}{k}: x\in F\},$2

including the clean identity

$\mathcal S_x=\{F\in \tbinom{[n]}{k}: x\in F\},$3

(Chen et al., 2019).

A further development is strong stability. If

$\mathcal S_x=\{F\in \tbinom{[n]}{k}: x\in F\},$4

then, under the hypotheses of the stability theorem in (Frankl et al., 2023), there exists a triple $\mathcal S_x=\{F\in \tbinom{[n]}{k}: x\in F\},$5 such that $\mathcal S_x=\{F\in \tbinom{[n]}{k}: x\in F\},$6 differs from a pure triangle family $\mathcal S_x=\{F\in \tbinom{[n]}{k}: x\in F\},$7 by quantitatively controlled error terms. This makes the triangle configuration the local model for near-extremal ordinary diversity (Frankl et al., 2023).

3. Weighted, higher-order, and operational variants

Several generalizations retain the same star-versus-spread interpretation while altering the objective.

Variant	Definition	Representative result
Ordinary diversity	$\mathcal S_x=\{F\in \tbinom{[n]}{k}: x\in F\},$8	For large $\mathcal S_x=\{F\in \tbinom{[n]}{k}: x\in F\},$9, the benchmark scale is $i\in[n]$0 (Kupavskii, 2017)
$i\in[n]$1-weighted diversity	$i\in[n]$2	For $i\in[n]$3 and $i\in[n]$4, $i\in[n]$5 uniquely maximizes $i\in[n]$6 (Frankl et al., 2023)
Double-diversity	$i\in[n]$7	For $i\in[n]$8, the Fano $i\in[n]$9-graph is the unique extremal family (Frankl et al., 2022)
Symmetric-difference spread	$\mathcal F$0	For $\mathcal F$1, $\mathcal F$2, stars maximize $\mathcal F$3 (Wang et al., 18 Jun 2026)

The $\mathcal F$4-weighted theory interpolates between size and diversity. When $\mathcal F$5, maximizing $\mathcal F$6 is just the Erdős–Ko–Rado problem. When $\mathcal F$7, it becomes ordinary diversity. For larger $\mathcal F$8, the penalty on high degree becomes stronger. One paper determines the maximal families for $\mathcal F$9 for large $i$0: stars are optimal for $i$1; “two out of three” configurations govern the interval $i$2; and Fano-plane-based families $i$3 and $i$4 govern $i$5. The same work records the corresponding growth-rate drops

$i$6

across these phases (Magnan et al., 2023).

Higher-order diversity replaces deletion of one vertex by deletion of several. For $i$7-diversity,

$i$8

where $i$9 is ordinary diversity and $\Delta(\mathcal F)=\max_{i\in[n]} d_i.$0 is double-diversity. The exact double-diversity theorem states that if $\Delta(\mathcal F)=\max_{i\in[n]} d_i.$1, then

$\Delta(\mathcal F)=\max_{i\in[n]} d_i.$2

with equality if and only if $\Delta(\mathcal F)=\max_{i\in[n]} d_i.$3 is isomorphic to the Fano $\Delta(\mathcal F)=\max_{i\in[n]} d_i.$4-graph. For triple diversity, the best stated result is an upper bound

$\Delta(\mathcal F)=\max_{i\in[n]} d_i.$5

and for $\Delta(\mathcal F)=\max_{i\in[n]} d_i.$6 the theory is much coarser (Frankl et al., 2022).

A related operational statistic is the symmetric-difference family $\Delta(\mathcal F)=\max_{i\in[n]} d_i.$7. Although it is not itself a diversity parameter, it is presented as a symmetric-difference analogue of classical extremal questions for intersecting families and is stated to be closely related to intersecting diversity. In the proved range $\Delta(\mathcal F)=\max_{i\in[n]} d_i.$8, $\Delta(\mathcal F)=\max_{i\in[n]} d_i.$9,

$$\gamma(\mathcal F)=|\mathcal F|-\Delta(\mathcal F),$$0

with equality only for stars (Wang et al., 18 Jun 2026).

4. Structural methods and extensions beyond set systems

The theory is driven by structural reduction. One influential large-$$\gamma(\mathcal F)=|\mathcal F|-\Delta(\mathcal F),$$1 approach uses the Dinur–Friedgut junta method: a sufficiently large intersecting family is essentially contained in a bounded-size junta, and Proposition 4 in (Kupavskii, 2017) gives a dichotomy for intersecting juntas that either places the family inside an $$\gamma(\mathcal F)=|\mathcal F|-\Delta(\mathcal F),$$2-type structure or forces diversity to be smaller than the benchmark. The exact extremal comparison is then completed by a cross-intersecting lemma proved with Kruskal–Katona and lexicographic compression (Kupavskii, 2017).

A different line uses shifting ad extremis. The 2023 improvement from $$\gamma(\mathcal F)=|\mathcal F|-\Delta(\mathcal F),$$3 to $$\gamma(\mathcal F)=|\mathcal F|-\Delta(\mathcal F),$$4 relies on shifting until only a small set of resistant pairs remains, followed by fiber-counting and cross-intersecting inequalities. A key structural claim is that the graph of shift-resistant pairs cannot contain three pairwise disjoint edges. This converts size information into a lower bound on $$\gamma(\mathcal F)=|\mathcal F|-\Delta(\mathcal F),$$5, and then into an upper bound on diversity (Frankl et al., 2023).

For $$\gamma(\mathcal F)=|\mathcal F|-\Delta(\mathcal F),$$6-weighted diversity, a variant of Frankl’s Delta-system method called the flower base plays the same role. A flower with threshold $$\gamma(\mathcal F)=|\mathcal F|-\Delta(\mathcal F),$$7 is a family with a common core $$\gamma(\mathcal F)=|\mathcal F|-\Delta(\mathcal F),$$8 whose petals outside $$\gamma(\mathcal F)=|\mathcal F|-\Delta(\mathcal F),$$9 have transversal number $\mathcal F\subseteq \binom{[n]}{k}$00, and the Flower Lemma states that every sufficiently large family of $\mathcal F\subseteq \binom{[n]}{k}$01-sets contains such a flower. The flower base $\mathcal F\subseteq \binom{[n]}{k}$02 is Sperner, intersecting when $\mathcal F\subseteq \binom{[n]}{k}$03 is intersecting, covers every edge of $\mathcal F\subseteq \binom{[n]}{k}$04, and has bounded size. This makes it possible to classify the small skeleton rather than the original family, producing the star / two-out-of-three / Fano-plane trichotomy (Magnan et al., 2023).

Cross-intersecting families admit a parallel diversity theory. In that setting, one writes the degree part and diversity part of each family as $\mathcal F\subseteq \binom{[n]}{k}$05. A 2026 structural theorem extends Kupavskii’s theorem from a single intersecting family to large cross-intersecting pairs and shows that, once the diversity parts are fixed, the maximal degree parts are the maximal cross-intersecting extensions. A technical innovation is the $\mathcal F\subseteq \binom{[n]}{k}$06-shift, designed to preserve both cross-intersection and local structural constraints (Huang et al., 18 Jun 2026).

The same philosophy has been exported to other combinatorial categories. For $\mathcal F\subseteq \binom{[n]}{k}$07-subspaces of $\mathcal F\subseteq \binom{[n]}{k}$08, diversity is the number of subspaces not containing the most popular $\mathcal F\subseteq \binom{[n]}{k}$09-dimensional subspace. In wide parameter regimes, the family $\mathcal F\subseteq \binom{[n]}{k}$10 has the largest diversity, with

$\mathcal F\subseteq \binom{[n]}{k}$11

and the same paper proves a Frankl-type degree-diversity theorem with extremal families $\mathcal F\subseteq \binom{[n]}{k}$12 (Ihringer et al., 4 May 2026). A complementary 2026 paper studies ordered point-degrees $\mathcal F\subseteq \binom{[n]}{k}$13 and proves, for intersecting families of $\mathcal F\subseteq \binom{[n]}{k}$14-subspaces with $\mathcal F\subseteq \binom{[n]}{k}$15,

$\mathcal F\subseteq \binom{[n]}{k}$16

together with a corrected $\mathcal F\subseteq \binom{[n]}{k}$17-analogue of the Huang–Rao $\mathcal F\subseteq \binom{[n]}{k}$18-th degree theorem for fixed $\mathcal F\subseteq \binom{[n]}{k}$19, sufficiently large $\mathcal F\subseteq \binom{[n]}{k}$20, and $\mathcal F\subseteq \binom{[n]}{k}$21 (Ma et al., 7 Jun 2026).

For permutation families $\mathcal F\subseteq \binom{[n]}{k}$22, intersecting means that every pair of permutations agrees in at least one position. Diversity is then the minimum number of permutations whose deletion results in a star. For $\mathcal F\subseteq \binom{[n]}{k}$23,

$\mathcal F\subseteq \binom{[n]}{k}$24

and this is sharp, with equality attained by triangle families. The proof uses the spread approximation method of Kupavskii and Zakharov together with Füredi’s pseudo-sunflower theorem (Wang et al., 12 Jan 2025).

5. A distinct probabilistic metric for intersecting demographic traits

A separate literature uses intersecting diversity to quantify aggregate identity composition in groups. Consider a group $\mathcal F\subseteq \binom{[n]}{k}$25 of $\mathcal F\subseteq \binom{[n]}{k}$26 individuals and $\mathcal F\subseteq \binom{[n]}{k}$27 traits. If the possible aggregate identities are

$\mathcal F\subseteq \binom{[n]}{k}$28

and $\mathcal F\subseteq \binom{[n]}{k}$29 is the proportion of the group with identity $\mathcal F\subseteq \binom{[n]}{k}$30, then intersecting diversity is defined by

$\mathcal F\subseteq \binom{[n]}{k}$31

where $\mathcal F\subseteq \binom{[n]}{k}$32 is the number of shared traits between two independently sampled individuals. Equivalently, $\mathcal F\subseteq \binom{[n]}{k}$33 is a normalized probability that two sampled individuals differ in at least one trait (Hoogstra et al., 11 Aug 2025).

The companion metric is shared identity,

$\mathcal F\subseteq \binom{[n]}{k}$34

The paper also defines $\mathcal F\subseteq \binom{[n]}{k}$35 using sampling without replacement and proves

$\mathcal F\subseteq \binom{[n]}{k}$36

Its main theorem establishes the bounds

$\mathcal F\subseteq \binom{[n]}{k}$37

the lower bound

$\mathcal F\subseteq \binom{[n]}{k}$38

and the gradient inequality

$\mathcal F\subseteq \binom{[n]}{k}$39

These formulas show that intersecting diversity and shared identity are structurally anti-correlated and that there is no clear “optimal” point maximizing both metrics simultaneously (Hoogstra et al., 11 Aug 2025).

The paper works out three case studies. In Hollywood-movie crews and Survivor tribes, the empirical $\mathcal F\subseteq \binom{[n]}{k}$40 clouds lie inside the admissible region and display the predicted anti-correlation. In Survivor, seasons beginning in 2020 are reported as more diverse overall than earlier seasons, while trait-by-trait analysis shows that race/ethnicity diversity increased and age diversity decreased. For North American companies, pairwise comparisons in which one company had both higher $\mathcal F\subseteq \binom{[n]}{k}$41 and higher $\mathcal F\subseteq \binom{[n]}{k}$42 yielded a higher industry-adjusted EBIT margin only $\mathcal F\subseteq \binom{[n]}{k}$43 of the time, and the paper states that this points in the opposite direction of the usual “more diversity + more shared identity = better performance” narrative (Hoogstra et al., 11 Aug 2025).

6. Broader intersectional and networked interpretations

Several adjacent literatures study intersecting forms of diversity without using the extremal-set-theoretic parameter. Information decomposition gives one such formulation. Using partial information decomposition, one paper interprets intersectional synergy as information about an outcome available only from the combination of identities, not from any identity alone. In U.S. census microdata, the relationship race + sex $\mathcal F\subseteq \binom{[n]}{k}$44 income is reported as about $\mathcal F\subseteq \binom{[n]}{k}$45 synergy, about $\mathcal F\subseteq \binom{[n]}{k}$46 redundancy, and about $\mathcal F\subseteq \binom{[n]}{k}$47 unique information total; the same paper uses synthetic data to show that linear regression with multiplicative interaction coefficients does not distinguish genuinely synergistic effects from redundant ones (Varley et al., 2021).

In HCI, type abstraction is used to avoid the combinatorial explosion of intersectional analysis. A formal compositional theorem states

$\mathcal F\subseteq \binom{[n]}{k}$48

so that separate analyses along different diversity dimensions can be joined by union. The claim is not that empirical work disappears, but that prior analytical artifacts can be reused when moving from one-dimensional to intersectional populations (Burnett et al., 2022).

Networked collective-learning models add a further layer. In one such model, diversity is implemented as heterogeneity in payoff functions. The main result is conditional: for simple tasks, diversity consistently impairs performance, whereas for complex tasks the effect depends on network density—diversity hurts in sparse networks and helps in dense networks, including the fully connected limit (Baumann et al., 2023). A related empirical study of everyday geography uses 49 mobility surveys, 385,000 respondents, and 1,711,000 trips to examine hourly intersectional urban patterns across gender, age, and education; it reports that in strongly daytime-attractive and strongly nighttime-decreasing districts, dominant groups are much more synchronous than non-dominant groups (Vallée et al., 2021).

A broader socio-technical synthesis appears in work on the “Internet of Us,” which models each participant by a vector of visible and invisible profile features and distinguishes descriptive diversity from prescriptive diversity. In that framework, diversity-aware AI is used to diversify rankings, mediate norms, and allow communities to choose which profile characteristics matter for diversification in their setting. This suggests a pragmatic, system-design interpretation of intersecting diversity as the joint management of multiple dimensions of difference rather than the optimization of a single scalar score (Michael et al., 17 Feb 2025).

Across these literatures, a common pattern persists. Intersecting diversity is not a synonym for size, heterogeneity, or fairness in the abstract. In extremal set theory it is a calibrated distance from a star; in the aggregate-identity metric it is a normalized probability of pairwise difference; in adjacent intersectional and networked work it becomes a way to formalize irreducible joint effects, compositional analysis, or context-dependent benefits of heterogeneous groups. The shared theme is that one-dimensional summaries are insufficient once intersection, concentration, or joint identity structure becomes the main object of study.