← The collection

Homophily.

Change same-group linking preferences and count actual cross-group connections.

Interactive experimentintuitiveField note ·
INTERACTIVE EXPERIMENT / P65

Homophily

Change same-group linking preferences and count actual cross-group connections.

TEST THE MECHANISM

Twelve nodes belong to two groups of six. Increase a same-group linking preference while holding the random draws fixed. Within- and cross-group connection probabilities change, so total edge count can change too.

01
123456789101112
Within-group connection probability
0.47
Cross-group connection probability
0.48
Actual within-group edges
11
Actual cross-group edges
19

Homophily describes similarity among connected people. This authored link rule supplies a preference mechanism, but an observed network can also reflect opportunities, institutions or shared geography; the drawing cannot identify the cause by itself.

The supplied linking rule is one possible mechanism, not an identified cause in real data. Total edge count can change; geography and institutions are omitted.

THE SHORT VERSION

Homophily, explained.

Homophily is the tendency for connected people to be similar on a specified characteristic, arising through preferences, opportunities or other mechanisms.

01 / THE MECHANISM

Why it happens

Twelve nodes form two groups. Same-group connection probability rises from 0.15 to 0.8 as the preference parameter grows, while cross-group probability falls from 0.8 to 0.15 under fixed random draws.

Homophily is the tendency for connected people to be similar on a specified characteristic, arising through preferences, opportunities or other mechanisms.

Read the result

Compare actual edge counts with the supplied connection probabilities. The total number of edges is not held fixed while the preference changes.

02 / FOLLOW IT THROUGH

A worked example

Change a mechanism that favours similar-group links

  1. Inspect the twelve nodes in two groups of six at an intermediate preference.

  2. Increase preference while preserving the random draws and compare within-group and cross-group links.

  3. Distinguish the stipulated preference mechanism from an inference about why a real observed network looks similar.

OPTIONAL DEEPER DETAILGo deeper: inside the model

Inside this model

Twelve nodes form two groups. Same-group connection probability rises from 0.15 to 0.8 as the preference parameter grows, while cross-group probability falls from 0.8 to 0.15 under fixed random draws.

03 / BEYOND THE EXPERIMENT

Where this idea is useful

A practical use

Distinguish an observed network pattern from an explanation based solely on personal preference.

CHECK YOUR INTUITION

A common misconception

THE TEMPTING CONCLUSION

“Similarity among friends must be caused by an active preference for similar people.”

THE MORE USEFUL DISTINCTION

Shared settings, geography and institutions can also shape opportunities for contact; a network pattern alone does not identify one mechanism.

What this explanation leaves out

  • The supplied linking rule is one possible mechanism, not an identified cause in real data. Total edge count can change; geography and institutions are omitted.
ONE MORE QUESTION

Why does the number of links change with the slider?

The model changes within- and cross-group probabilities rather than rewiring a fixed number of edges.

TAKE THE IDEA WITH YOU

What opportunities to meet were available before you attributed a network pattern to preference?

Associated thinkers

Associated withMiller McPherson ↗

Associations marked provisional are awaiting source review.

Further reading

Explore the original research or the teaching reference behind this experiment.