Application of Hierarchical/Multilevel Models and Quality of Reporting (2010–2020): A Systematic Review

Table 2

Characteristics of statistical inference and estimation methods.

Method of estimation
Bayesian	13 (20.0%)
GEE	1 (1.5%)
GLS	2 (3.1%)
MLE	12 (18.5%)
PQL	1 (1.5%)
ReML	3 (4.6%)
ReML and Bayesian	1 (1.5%)
ReML and MLE	3 (4.6%)
Weighted least squares	2 (3.1%)
NI	27 (41.5%)
Distribution of response
Binary	24 (36.9%)
Multinomial	1 (1.5%)
Normal	31 (47.7%)
Normal and binary	2 (3.1%)
Ordinal and binary	1 (1.5%)
Ordinal	5 (7.7%)
Poisson	1 (1.5%)
Model validation measures
AIC	3 (4.6%)
BIC	2 (3.1%)
AIC and BIC	7 (10.8%)
AIC, BIC, and LRT	11 (16.9%)
DIC	7 (10.8%)
R-square	3 (4.6%)
LRT	5 (7.7%)
RMSEA/MSE	6 (9.2%)
HDI/QIC	2 (3.1%)
NI	19 (29.2%)
Statistical software
HLM	6 (9.2%)
Mlwin	8 (12.3%)
Mplus	4 (6.2%)
R	13 (20.0%)
SAS	9 (13.8%)
SAS and Mlwin	1 (1.5%)
SPSS	5 (7.7%)
Stata	9 (13.8%)
Stata and Mlwin	2 (3.1%)
WinBUGS	2 (3.1%)
NI	6 (9.2%)
Statistical packages
GLIMMIX	5 (7.7%)
GLLAMM	1 (1.5%)
HSAR	1 (1.5%)
JAGS (rjags, runjags, and coda)	2 (3.1%)
lme4/lmer	4 (6.2%)
lme4, spdep, and sjstats	2 (3.1%)
Nmle	2 (3.1%)
PROC MIXED/NLMIXED	4 (6.2%)
VCMMR estimation	1 (1.5%)
xtlogit	1 (1.5%)
xtmixed	1 (1.5%)
NI	41 (63.1%)