Reducing the bias due to unknown relationships in measuring the steepness of a dominance hierarchy

Measuring the steepness of a dominance hierarchy is important for classifying social system in continuum between egalitarian and despotic. For this, often steepness-slope from de Vries et al. (Animal Behaviour, 2006, 71, 585–592) used. It compares cardinal ordinal rank each individual using slope linear regression. The disadvantage this measure that becomes lower higher proportion unknown relationships (dyads without interactions). In present paper, we investigate what causes bias, propose solution. (1) We show bias due to treatment by index currently used methodology, David's score (namely assuming, among other things, an equal number wins losses both members pair). (2) Instead, Average Dominance Index (the average all its opponents) reduces relationships, because it excludes these (3) standard error based on smaller. (4) two indices (David's Index) result similar steepness-slopes when are known. To compare use empirical data (from four group-years wild vervet monkeys, Chlorocebus pygerythrus) computational model interactions group (DomWorld). conclude (compared score) preferable measuring steepness-slope.