Standardized effect inference
In comparing my reaction times with those of my son Steven, we are interested in the regression model
$$ Time = \beta_0 + \beta_1 * Person $$
The parameter beta1
measures the change (from me to Steven) in mean reaction time and the parameter
delta = beta1 / S
represents the standardized change (in standard deviation units).
This exercise is part of the course
Beginning Bayes in R
Exercise instructions
- Perform a regression fit of
Time
usingPerson
as a covariate. - Simulate 1000 draws from the joint posterior distribution using the
sim()
function. - Extract the simulated values of (
beta0, beta1
) andS
(functionscoef()
andsigma.hat()
). - Compute the simulated values of the standardized change
delta
. - Find a 90% interval estimate for
delta
.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Perform a regression fit of Time with Person as a covariate: fit
fit <- lm(Time ~ Person, data = dd)
# Simulate 1000 values from the posterior distribution: sim_fit
sim_fit <- sim(fit, n.sims = 1000)
# Extract simulated draws of beta and S: beta_sim, s_sim
beta_sim <- ___
s_sim <- ___
# Compute simulated values of the standardized change: s_delta
# Find 90% interval estimate for s_delta