get_individual_coefficients: Calculate Individual-Level Predicted Coefficients from...

In beezdemand: Behavioral Economic Easy Demand

Calculate Individual-Level Predicted Coefficients from beezdemand_nlme Model

Description

This function extracts and combines fixed and random effects to calculate individual-level predicted coefficients for all parameter-factor combinations from a beezdemand_nlme model object. It automatically detects the factor structure and calculates coefficients for each individual and factor level.

Usage

get_individual_coefficients(
  fit_obj,
  params = c("Q0", "alpha"),
  format = c("wide", "long")
)

Arguments

fit_obj	A beezdemand_nlme object returned by fit_demand_mixed().
params	Character vector specifying which parameters to calculate. Options are "Q0", "alpha", or c("Q0", "alpha"). Default is c("Q0", "alpha").
format	Character, output format. "wide" returns one row per individual with separate columns for each parameter-factor combination. "long" returns one row per individual-parameter-factor combination. Default is "wide".

Details

Individual-level coefficients represent the predicted parameter values for each subject in the study. For models with factors, these coefficients combine:

1. The baseline intercept effect (fixed + random)

2. The factor-specific effect (fixed + random) for each factor level

This is equivalent to manually calculating: coefficient = intercept_fixed + intercept_random + factor_fixed + factor_random

The function automatically handles:

• Models with or without factors

• Any number of factor levels

• Missing random effects (defaults to 0)

• Complex factor structures with multiple factors

For models without factors, only intercept coefficients are calculated. For models with factors, both intercept and factor-level coefficients are provided.

Value

A data frame with individual-level predicted coefficients.

• In "wide" format: rows are individuals, columns are parameter-factor combinations

• In "long" format: columns are id, parameter, condition, coefficient_value

Column naming convention for wide format:

• ⁠estimated_\{param\}_intercept⁠: Baseline/reference level coefficient

• ⁠estimated_\{param\}_\{factor\}\{level\}⁠: Factor level-specific coefficient

All coefficients are on the log10 scale (same as model estimation scale). To convert to natural scale, use 10^coefficient.

See Also

fit_demand_mixed for fitting the original model coef.beezdemand_nlme for extracting model coefficients get_demand_param_emms for estimated marginal means

Examples

data(ko)
fit <- fit_demand_mixed(ko, y_var = "y_ll4", x_var = "x",
                        id_var = "monkey", factors = "drug",
                        equation_form = "zben")
individual_coefs <- get_individual_coefficients(fit)
head(individual_coefs)

beezdemand documentation built on March 3, 2026, 9:07 a.m.