Where does tie collide?

In rank-based statistics a tie occurs when two or more observations share the same value, so they cannot be placed in a strict order and instead share the average of the ranks they would have taken. Spearman's 1904 The Proof and Measurement of Association Between Two Things introduced correlation computed from ranks, where equal values are the case that has to be resolved this way. Kendall's 1938 A New Measure of Rank Correlation defined a coefficient with a correction that adjusts for the number of tied pairs. This referent is a condition in a dataset, not a member under load, a bond between people, or a mark in a score.[1][2]

Unit

A group of observations holding one equal value, counted by how many share it

Recovery means

Not applicable

Measured by

Handled by assigning mean ranks and applying tie corrections in coefficients such as Kendall's tau-b and Spearman's rho

In the study of social networks a tie is a specific relationship that connects two actors, such as friendship, advice, kinship, or the passing of information, and it is the basic link from which a network is built. Moreno's 1934 Who Shall Survive? mapped such relationships as lines between people in the sociogram, founding sociometry. Granovetter's 1973 The Strength of Weak Ties made the tie and its strength a central analytic unit, showing that weaker acquaintance ties often carry information a person's close circle does not. This referent is a bond between people, not a member in tension, a data condition, or a notated duration.[3][4]

Unit

A single relationship between two actors, sometimes weighted by its strength or frequency

Recovery means

Not applicable

Measured by

Recorded by survey name generators, observation, or archival records, then represented as edges in a network graph