topic_model_data: Topic model dataset

In MLBC: Bias Correction Methods for Models Using Synthetic Data

Topic model dataset

Description

Dataset containing topic model outputs for demonstrating bias correction methods in topic model regressions using CEO diary data.

Usage

topic_model_data

Format

A list with 8 components:

covars

Data frame (916 x 11): Control variables

estimation_data

Data frame (916 x 672): Contains outcome ly and word frequencies

gamma_draws

Data frame (2000 x 2): MCMC draws

theta_est_full

Data frame (916 x 2): Full sample topic proportions

theta_est_samp

Data frame (916 x 2): Subsample topic proportions

beta_est_full

Data frame (2 x 654): Full sample topic-word distributions

beta_est_samp

Data frame (2 x 654): Subsample topic-word distributions

lda_data

Data frame (916 x 2): LDA validation data

Source

CEO diary data from Bandiera et al (2020), Journal of Political Economy

See Also

ols_bca_topic, ols_bcm_topic

Examples

data(topic_model_data)

# Basic exploration
Y <- topic_model_data$estimation_data$ly
theta <- as.matrix(topic_model_data$theta_est_full)

cat("Sample size:", length(Y), "\n")
cat("Mean log employment:", round(mean(Y), 2), "\n")
cat("Topic 1 mean:", round(mean(theta[, 1]), 3), "\n")

MLBC documentation built on Aug. 8, 2025, 7:31 p.m.