Your friends have more friends than you
It is not an impression, it is a theorem. Proved in 1991, it follows from how a network is sampled when one follows its ties.
The statement
Scott Feld published in 1991 an article showing that, in almost any social network, the average number of relationships of the people someone names exceeds the average number of relationships in the population. The article appeared in the American Journal of Sociology under a title that sums up the result: why your friends have more friends than you do.
The result is not an empirical observation but a mathematical property of graphs: it holds as soon as degrees are not all identical. The gap grows with the dispersion of the numbers of relationships: in a network where a few people have a great many, it becomes considerable.
Why it is true
A person with many relationships appears in many lists, a person with few appears in few. Drawing a friend at random therefore amounts to drawing a person in proportion to their number of relationships.
It is the same mechanism that makes a randomly drawn train passenger find carriages fuller than the average carriage: one samples seats, not trains.
The paradox holds on average, not for everyone. A very well-connected person may have friends less connected than they are; only the average computed over the whole network obeys the theorem.
What it implies for surveys
Any method recruiting participants by following a network's ties overrepresents highly connected people, and produces biased estimates if correction is not applied.
That property is also used deliberately: following ties reaches central people faster, a method used in epidemiology for early detection of spread. Nicholas Christakis and James Fowler applied it in 2010 to a flu outbreak on a campus: the friends named by randomly selected students were affected earlier than those students themselves.
The subjective side
The mechanism applies to any characteristic correlated with the number of relationships. Highly connected people go out more often, which explains part of the sense that others have a denser social life.
Studies have extended that reasoning to other attributes, notably on platforms, where the most visible content comes from the most active accounts, hence the least representative. The same bias thus shapes the picture each person forms of other people's sociability, without anyone being mistaken about what they observe.
What it does not say
It says nothing about the quality of relationships, nor about well-being. It is a sampling property, and interpreting it as a social finding is a frequent error. Having fewer relationships than one's friends is the most common situation, and says nothing about the person who notices it.
Its value is methodological: it recalls that an observation made from inside a network is never a representative sample of that network.
Key points
The 1991 result is a mathematical property, true as soon as degrees differ.
Following a network's ties mechanically overrepresents highly connected people.
The property extends to any attribute correlated with the number of relationships, including online visibility.
This article describes documented mechanisms and disagreements. It recommends no position, no party and no vote, and commits to no reading other than that of the work it cites.
And you, where do you stand?
The Civix test situates your convictions on six dimensions and matches you with one of the twelve profiles, without ever telling you what to think.
Start my test