A comparative review of methods for comparing means using partially paired data

  title={A comparative review of methods for comparing means using partially paired data},
  author={Beibei Guo and Ying Yuan},
  journal={Statistical Methods in Medical Research},
  pages={1323 - 1340}
In medical experiments with the objective of testing the equality of two means, data are often partially paired by design or because of missing data. The partially paired data represent a combination of paired and unpaired observations. In this article, we review and compare nine methods for analyzing partially paired data, including the two-sample t-test, paired t-test, corrected z-test, weighted t-test, pooled t-test, optimal pooled t-test, multiple imputation method, mixed model approach… 

Figures and Tables from this paper

Analyzing partially paired data: when can the unpaired portion(s) be safely ignored?
The findings are counterintuitive, i.e. the ‘P-value pooling methods’ do not always outperform the naive paired tests in terms of power and this paper presents guidance on how to select the best test for testing equality of means with partially paired data.
Optimal weighted two-sample t-test with partially paired data in a unified framework
It is found that the proposed test statistic is generally more powerful than existing t-tests and can be viewed as special members in this unified framework.
Hypothesis testing for matched pairs with missing data by maximum mean discrepancy: An application to continuous glucose monitoring
A frequent problem in statistical science is how to properly handle missing data in matched paired observations. There is a large body of literature coping with the univariate case. Yet, the ongoing
Parametric and non-parametric tests for the comparison of two samples
  • Mathematics
  • 2018
Basic teaching of statistics usually assumes a perfect world with completely independent samples or completely dependent samples, examples include Swinscow (2002) and OpenStax (2013). Real world
Testing equality of means in partially paired data with incompleteness in single response
This paper aims to fill the gap on paired data with incompleteness in single arm testing by proposing some new methods, namely, P-value pooling methods and a nonparametric combination test.
Testing marginal homogeneity of a continuous bivariate distribution with possibly incomplete paired data
We discuss the testing problem of homogeneity of the marginal distributions of a continuous bivariate distribution based on a paired sample with possibly missing components (missing completely at
Review of the partially overlapping samples framework: Paired observations and independent observations in two samples
A frequently asked question in quantitative research is how to compare two samples that include some combination of paired observations and unpaired observations. In our publications and R package,
Permuting incomplete paired data: a novel exact and asymptotic correct randomization test
ABSTRACT Various statistical tests have been developed for testing the equality of means in matched pairs with missing values. However, most existing methods are commonly based on certain
Parametric and Non-Parametric Tests for the Comparison of Two Samples Which Both Include Paired and Unpaired Observations
Samples that include both independent and paired observations cause a dilemma for researchers that covers the full breadth of empirical research. Parametric approaches for the comparison of two
Much Ado about Almost Nothing: Methods for Dealing with Limited Data
: Applied statisticians are often confronted with statistical inference problems dealing with situations in which there appear to be no data, or data of only limited usefulness. For example, when


On comparing means in the paired case with incomplete data on both responses
SUMMARY Five procedures for testing the equality of two correlated means with incomplete data on both responses are compared by using Monte Carlo studies; three procedures were suggested by Lin &
Notes on two sample tests for partially correlated (paired) data
We provide several methods to compare two Gaussian distributed means in the two sample location problems under the assumption of partially dependent observations. Simulation studies indicate that our
A method for comparing two normal means using combined samples of correlated and uncorrelated data.
The simulation results indicate that the proposed method can provide substantial improvement in type I error rate and power when compared with standard methods of analysis.
Using the general linear mixed model to analyse unbalanced repeated measures and longitudinal data.
The purpose of this tutorial is to provide readers with a sufficient introduction to the theory to understand the method and a more extensive discussion of model fitting and checking in order to provide guidelines for its use.
Using the General Linear Mixed Model to Analyse Unbalanced Repeated Measures and Longitudinal Data
The general linear mixed model provides a useful approach for analysing a wide variety of data structures which practising statisticians often encounter. Two such data structures which can be
Variance of a Weighted Mean
In many areas of statistical practice the problem arises of combining several estimates of an unknown quantity to obtain an estimate of improved precision. For example, suppose two or more
Random-effects models for longitudinal data.
A unified approach to fitting these models, based on a combination of empirical Bayes and maximum likelihood estimation of model parameters and using the EM algorithm, is discussed.
On difference of means with incomplete data
SUMMARY An estimate of the difference of means is obtained when sampling from a bivariate normal distribution with variances o-2 and o- and correlation p, where some observations on either of the
Small-sample degrees of freedom with multiple imputation
An appealing feature of multiple imputation is the simplicity of the rules for combining the multiple complete-data inferences into a final inference, the repeated-imputation inference (Rubin, 1987).