Statistics Glossary → Ties method (Efron, Breslow, exact)
Ties method (Efron, Breslow, exact)concept
The method that determines how the partial likelihood of Cox regression is computed when several events occur at the same time (tied times).
The partial likelihood of the Cox model assumes that events are ordered one by one in time. When times are recorded in coarse units such as days or months, several events occur at the same moment and their order is unknown; these are called ties.
Breslow's method is the simplest approximation and pulls coefficients towards zero when ties are many; Efron's method is more accurate and is R's default; the exact method accounts for all possible orderings but is slow. With few ties the three give almost the same result.
An everyday example
Suppose finishing times in a race were recorded only in whole minutes; three runners appear to have finished 'in minute 12', but their order is unknown. Any calculation based on the ranking needs a rule for how to treat these three runners.
In MerQur
The 'Ties method (Efron / Breslow / Exact)' option of Cox regression; MerQur's exact option is the discrete-time (conditional logistic) exact likelihood.
Analyses where this concept appears
Cox Regression (Proportional Hazards Model)