rep_consensus: Reputation Consensus Model

In Coryc3133/ReputationAnalyses: Functions to model and analyze triadic (T-P1-P2) reputational data

Description

This fits a model estimating the possible hearsay reputation parameters from a combination of P1- and P2-reports (no self-reports or accuracy criterion). It requires a dataframe and either a model from the relevant model builder function or names of columns with P1- and P2- ratings. The estimated parameters are:

hc

hearsay consensus; the correlation between P1(T) & P2(T)

int_p1

Intercept for P1(T)

int_p2

Intercept for P2(T)

v_p1

variance for P1(T)

v_p2

variance for P2(T)

p1_p2_rel_el

P1-P2 Relative Elevation (i.e., Mean P1(T) - Mean P2(T))

If n exchangeable triads > 1:

rec

direct reciprocity; the correlation between opposit P1(T)s (e.g., A(C) <-> C(A))

h

hearsay reciprocity; the correlation between exchangeable P2(T)s (e.g., B(C) <-> D(A))

m

unnamed parameter; The correlation between P2(T) and the opposite P1(T) in a group. (e.g., B(C) <-> C(A))

The function can handle up to n exchangeable triads.

Usage

rep_consensus(data, model = NULL, p1_reports, p2_reports,
  n_triads = length(p1_reports), n_p1s_per_p2s = 1,
  n_p2s_per_p1s = 1)

Arguments

data	The dataframe that contains P1 and P2 ratings. Data should be wide, with a row for every group of participants. At a minimum, it must contain two columns: one for P1 reports and one for P2 reports.
model	Optional. A model from the corresponding ReputationModelR model builder function. If this is supplied, no additional arguments need to be specified.
p1_reports	Quoted column names that contain P1 reports, or ratings made by the person that knows the target directly. If more than one is supplied, the target-wise order must match the other rating types.
p2_reports	Quoted column names that contain P2 reports, or ratings made by the person that knows the target indirectly through the corresponding P1. If more than one is supplied, the target-wise order must match the other rating types.
n_triads	The number of exchangeable triads in each group. By default, this is determined by counting the number of P1 reports. This parameter rarely needs to be changed.
n_p1s_per_p2s	The number of P1s for every P2. This defaults to 1. Currently, only values of 1 are supported.
n_p2s_per_p1s	The number of P2s for every P1;. This defaults to 1. Currently, only values of 1 are supported.

Value

The function returns an object of class lavaan.

Examples

data("rep_sim_data")
          agree_consensus <- rep_consensus(data = rep_sim_data,
                                           p1_reports = c("A_C_agreeableness", "C_A_agreeableness"),
                                           p2_reports = c("B_C_agreeableness", "D_A_agreeableness"))
         # alternatively
         # build the model
          agree_consensus_model <- rep_consensus_builder(p1_reports = c("A_C_agreeableness", "C_A_agreeableness"),
                                                             p2_reports = c("B_C_agreeableness", "D_A_agreeableness"))
         # then fit it
         agree_consensus <- rep_consensus(data = rep_sim_data,
                                          agree_consensus_model)

Coryc3133/ReputationAnalyses documentation built on July 31, 2019, 8:35 a.m.