Simulation Results
Format
A dataframe with 24,544 rows and 13 columns:
- zero_hit
The proportion of replications where the confidence intervals contained zero.
- theta_hit
The proportion of replications where the confidence intervals contained the population \(\alpha \beta\).
- replications
Simulation replications.
- taskid
Simulation Task ID.
- tauprime
\(\tau^{\prime}\), that is, the path from \(X\) to \(Y\), adjusting for \(M\).
- beta
\(\beta\), that is, the path from \(M\) to \(Y\).
- alpha
\(\alpha\), that is, the path from \(X\) to \(M\).
- n
Sample size.
- sigmasqepsilonm
Error variance \(\sigma^{2}_{\varepsilon_{M}}\).
- sigmasqepsilony
Error variance \(\sigma^{2}_{\varepsilon_{Y}}\).
- alphabeta
\(\alpha \beta\), that is, the indirect effect of \(X\) on \(Y\) via \(M\).
- mechanism
Missing data mechanism.
"COMPLETE"for complete data,"MCAR"for missing completely at random, and"MAR"for missing at random.- proportion
Proportion of missing data (.0, .1, .2, .3).
- method
Method used.
- type1
Type I error rate.
- power
Statistical power.
- miss
Miss rate.
The methods are as follows:
- MC.COMPLETE
for Monte Carlo method with maximum likelihood estimates for complete data.
- MC.FIML
for Monte Carlo method with full information maximum likelihood estimates.
- MC.MI
for Monte Carlo method with multiple imputation estimates.
- MC.MI.ADJ
for Monte Carlo method with adjusted multiple imputation estimates.
- NBBC.COMPLETE
for bias-corrected nonparametric bootstrap with maximum likelihood estimates for complete data.
- NBBC.FIML
for full maximum likelihood nested within bias-corrected nonparametric bootstrap.
- NBPC.COMPLETE
for percentile nonparametric bootstrap with full maximum likelihood estimates for complete.
- NBPC.FIML
for full maximum likelihood nested within percentile nonparametric bootstrap.
- SIG.COMPLETE
for joint-significant test for complete data.
- SIG.FIML
for the joint-significant test with full maximum likelihood estimates.
- SIG.MI
for joint-significant test with multiple imputation estimates.
- SIG.MI.ADJ
for the joint-significant test with adjusted multiple imputation estimates.