I have a simple mediation model:

IV: Experience (binary – yes1/no0)
Mediator: Trust (5 point likert scale measure)
DV: Offer (continuous, values can be anything between 5 and 15)

Each respondent (ID) was exposed to a scenario without experience and with experience. For each scenario they got a value for trust and for offer

My data looks something like this:

I want to conduct a mediation analysis in R studio, but I need to account for within subjects repeated measures (per ID).

WHAT I HAVE SO FAR:

install.packages("readxl")
install.packages("lavaan")
install.packages("sandwich")
install.packages("lmtest")
install.packages("dplyr")
install.packages("car")

library(readxl)
library(lavaan)
library(sandwich)
library(lmtest)
library(dplyr)
library(car)

# Load your data using the specific path
data <- read_excel("/Users/reka/Documents/RSM 2022-2023/THESIS/Main Data/Data.xlsx")

# Display the first few rows and column names to verify
head(data)
str(data)

# Convert variables to numeric
data$Time <- as.numeric(as.factor(data$Time)) # IV
data$Matrix <- as.numeric(as.factor(data$Matrix)) # Mediator
data$Offer <- as.numeric(data$Offer) # DV
data$ID <- as.factor(data$ID) # Cluster ID

# Model specification
model <- '
  # Direct effect of Time on Offer
  Offer ~ Time_to_Offer*Time
  # Effect of Time on Matrix (Mediator)
  Matrix ~ Time_to_Matrix*Time
  # Effect of Matrix on Offer
  Offer ~ Matrix_to_Offer*Matrix
  # Indirect effect
  indirect := Time_to_Matrix * Matrix_to_Offer
  # Direct effect
  direct := Time_to_Offer
  # Total effect
  total := direct + indirect
'

# Fit the mediation model using lavaan
fit <- sem(model, data = data)
summary(fit, standardized = TRUE)

After this I don’t know how to proceed to make sure i cluster standard errors by ID.

Is my mediation model even good?