rep_accuracy_id_mods: Individual level Moderators of Hearsay Accuracy

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

Description

This function fits a model for individual-level moderator on hearsay accuracy. It requires a dataframe and either a model from the relevant model builder function or names of columns for self-reports, p2_reports, and id_mod_variable. The estimated parameters are:

Usage

rep_accuracy_id_mods(data, model = NULL, target_self, p2_reports,
  id_mod_variable, interaction_term, n_triads = length(target_self),
  n_ts_per_p2s = 1, n_p2s_per_ts = 1)

Arguments

data	The dataframe that contains the ratings, moderator variable, and the interaction term. Data should be wide, with a row for every group of participants. At a minimum, it must contain four columns: one for target self-reports, one for mean-centered P2 reports, one for the mean-centered moderator variable, and one for the interaction term.
model	Optional. A model from the corresponding ReputationModelR model builder function. If this is supplied, no additional arguments need to be specified.
target_self	Quoted column names that contain target self-reports. If more than one is supplied, the 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. Ratings should be grand-mean-centered to increase the interpretibility of the model parameters. If more than one is supplied, the target-wise order must match the other rating types.
id_mod_variable	Quoted column names that contain the individual-level moderator of interest. If more than one is supplied from multiple exchangeable triads, the order must match the order of the ratings. Like P2-reports, the variable should be mean-centered to facilitate interpretability.
interaction_term	Quoted column names that contain the interaction term, or the product of the mean-centered P2-report and the mean-centered moderator variable. If more than one is supplied from multiple exchangeable triads, the target-wise order must match the order of the ratings.
n_triads	The number of exchangeable triads in each group. By default, this is determined by counting the number of target self-reports. This parameter rarely needs to be changed.
n_ts_per_p2s	The number of targets that each P2 rated. This defaults to 1. Currently, only values of 1 are supported.
n_p2s_per_ts	The number of P2s that rated each target;. This defaults to 1. Currently, only values of 1 are supported.

Details

The parameters for the individual-level moderator analyses are:

ha_me

hearsay accuracy main effect; this should correspond to hearsay accuracy at average level of moderator variable (if data were properly mean-centered).

mod_me

The meain effect of the moderator variable; it can be interpreted as the difference in target self-reports related to differences in the individual-level moderator variable.

interaction

This is the interaction term. It indicates the extent to which hearsay accuracy, depends on the moderator variable

v_self

variance for T(T)

v_p2

variance for P2(T)

v_mod

variance for moderator variable

v_interaction

variance for interaction term

int_self

intercept for T(T)

int_p2

intercept for P2(T)

int_mod

intercept for moderator variable

int_interaction

intercept for interaction term

The function can handle up to n exchangeable triads.

Value

The function returns an object of class lavaan.

Examples

data("rep_sim_data")
          # Prepare data
         moderator_data <- rep_sim_data %>%
           dplyr::mutate(B_C_agreeableness_cent = scale(B_C_agreeableness, scale = FALSE),
                  D_A_agreeableness_cent = scale(D_A_agreeableness, scale = FALSE),
                  B_iri_perspective_cent = scale(B_iri_perspective, scale = FALSE),
                  D_iri_perspective_cent = scale(D_iri_perspective, scale = FALSE),
                  B_ptXagree_interaction = B_C_agreeableness_cent*B_iri_perspective_cent,
                  D_ptXagree_interaction = D_A_agreeableness_cent*D_iri_perspective_cent)

# Fitting by supplying variable/column names
agree_pt_mods_hearacc <- rep_accuracy_id_mods(data = moderator_data,
                                target_self = c("C_C_agreeableness", "A_A_agreeableness"),
                                p2_reports = c("B_C_agreeableness_cent", "D_A_agreeableness_cent"),
                                id_mod_variable = c("B_iri_perspective_cent", "D_iri_perspective_cent"),
                                interaction_term = c("B_ptXagree_interaction", "D_ptXagree_interaction"))

# Alternatively, first build the model
agree_pt_mods_hearacc_model <- rep_accuracy_id_mods_builder(target_self = c("C_C_agreeableness", "A_A_agreeableness"),
                                p2_reports = c("B_C_agreeableness_cent", "D_A_agreeableness_cent"),
                                id_mod_variable = c("B_iri_perspective_cent", "D_iri_perspective_cent"),
                                interaction_term = c("B_ptXagree_interaction", "D_ptXagree_interaction"))

# Then fit the model you just built
 agree_pt_mods_hearacc <- rep_accuracy_id_mods(data = moderator_data,
                                               model = agree_pt_mods_hearacc_model)

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