Principled missing data methods for researchers

By Yiran Dong, C. Y. Peng. SpringerPlus, 2013

2,108 citations recorded by the source corpus.

The impact of missing data on quantitative research can be serious, leading to biased estimates of parameters, loss of information, decreased statistical power, increased standard errors, and weakened generalizability of findings. In this paper, we discussed and demonstrated three principled missing data methods: multiple imputation, full information maximum likelihood, and expectation-maximization algorithm, applied to a real-world data set. Results were contrasted with those obtained from the complete data set and from the listwise deletion method. The relative merits of each method are noted, along with common features they share. The paper concludes with an emphasis on the importance of statistical assumptions, and recommendations for researchers. Quality of research will be enhanced if (a) researchers explicitly acknowledge missing data problems and the conditions under which they occurred, (b) principled methods are employed to handle missing data, and (c) the appropriate treatment of missing data is incorporated into review standards of manuscripts submitted for publication.

Subfields: Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/9ac8ef178c71a93570ee44af6542983fd74ecb08

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

The Development and Validation of the Game User Experience Satisfaction Scale (GUESS) (2016, positive): “In general, research has shown that ML-based methods (e.g., EM, FIML) produce similar results (Dong & Peng, 2013; Graham, Olcho-wski, & Gilreath, 2007; Schafer & Graham, 2002).”
A survey on missing data in machine learning (2021, background): “The monotone data pattern is easier to deal with since patterns among the missing values are easily observable [24].”
An Overview of IoT Sensor Data Processing, Fusion, and Analysis Techniques (2020, methodology): “Thus, the raw sensor data need to undergo data cleaning processing, and then data analysis can be performed to obtain relevant information from this cleaned IoT sensor data [8,9].”
Multiple minority stress: The role of proximal and distal stress on mental health outcomes among lesbian, gay, and bisexual people of color (2019, result): “It is suggested that a missing data rate of less than 5–10% is inconsequential and does not bias the results (Bennett, 2001; Dong & Peng, 2013; Schafer, 1999), and that patterns of missing data have more of an influence on results than the proportion of missing data (Tabachnick & Fidell, 2013).”
The proportion of missing data should not be used to guide decisions on multiple imputation (2019, background): “Statistical guidance articles have stated that bias is likely in analyses with more than 10% missingness and that if more than 40% data are missing in important variables then results should only be considered as hypothesis generating [18,19].”
Loneliness in the UK during the COVID-19 pandemic: Cross-sectional results from the COVID-19 Psychological Wellbeing Study (2020, positive): “As less than 5% of the data were missing pairwise and listwise deletion were implemented [69,70].”
Effects of support on stress and burnout in school principals (2016, methodology): “Missing data were handled using full information maximum likelihood estimation (Enders 2010), which is a common method in structural equation modelling (Dong and Peng 2013).”