Table 2.

Structure of all the models used in the analysis.

sex-top-associates model (Model 1a)sex-significant-associate model (Model 1b)what-makes-top-associates model (Model 2a)what-makes-significant-associates model (Model 2b)association-skew model (Model 3)
responsesex of the topsex of theWas individual 1 topWas individual 1skew index of the PAV
associatesignificantassociate ofsignificant associatedistribution (SKPAV)
associatesindividual 2 (Y/N)of individual 2 (Y/N)
fixed factorsa
  1. sex of the individual

  2. species (chimpanzee/bonobo)

  3. species × sex

  1. sex of individual 1

  2. sex of individual 2

  3. species (bonobo/chimpanzee)

  4. Are individuals 1 and 2 kin? (N/Y)

  5. Were individuals 1 and 2 top/significant associate in the previous quarter? (N/Y)

  6. Sex 1 × Sex 2 × species

  7. kin × species

  8. top/significant associate in the previous quarter × species

  1. sex of individual 1

  2. sex of individual 2

  3. species (bonobo/chimpanzee)

  4. Sex 1 × Sex 2 × species

random intercepts
  1. individual ID

  2. community ID

  3. quarter

  1. individual 1 ID

  2. individual 2 ID

  3. dyad individual 1–individual 2

  4. community ID

  1. individual 1 ID

  2. quarter ID

  3. community ID

  4. quarter | individual 1 ID

random slopesb
  1. sex | community ID

  2. sex | quarter

  1. kinship | individual 1 ID

  2. kinship | individual 2 ID

  3. kinship | community ID

  4. sex individual 1 | individual 2 ID

  5. sex individual 1 | community ID

  6. sex individual 2 | individual 1 ID

  7. sex individual 2 | community ID

  8. top/significant associate in the previous quarter | individual 1 ID

  9. top/significant associate in the previous quarter | individual 2 ID

  10. top/significant associate in the previous quarter | community ID

  1. sex individual 1 | quarter

  2. sex individual 1 | community ID

  3. sex individual 2 | quarter

  4. sex individual 2 | community ID

  5. sex individual 2 | individual 1 ID

  6. Sex 1 × Sex 2 | community ID

  7. Sex 1 × Sex 2 | quarter

offsetproportion of males in the community
  • aSex was dummy coded with females being the reference category; species was dummy coded with bonobo being the reference category; kinship was dummy coded with non-kin being the reference; and ‘top/significant associate in the previous quarter’ was dummy coded with no being the reference category.

  • bFor inclusion as random slopes we manually dummy coded and then centred (to a mean of zero) kinship, sex and ‘top/significant associate in the previous quarter’.