Accounting for measurement error: a critical but often overlooked process

Area of Science:

• Scientific methodology
• Biostatistics
• Measurement science

Background:

• Measurement error, including technical error of measurement (TEM), is inherent in data collection due to instrument and human factors.
• Substantial TEM can significantly impact research results and statistical inferences across various scientific disciplines.
• Minimizing TEM is a crucial goal for robust data collection regimens.

Purpose of the Study:

• To review and discuss statistical approaches for identifying and controlling TEM.
• To guide researchers in distinguishing true experimental differences from measurement variability.
• To highlight the importance of accounting for error in statistical analysis.

Main Methods:

• Review of statistical designs related to types of error and error accountability.
• Discussion of landmark location, validity, technical error, systematic error, analysis of variance, scaled measures, and correlation coefficients.
• Emphasis on the role of repeated measurements in quantifying and controlling TEM.

Main Results:

• TEM is a component of unexplained variance in statistical tests.
• Accounting for TEM requires the incorporation of repeated measurements into study designs.
• Proper statistical accounting for TEM enhances the power to detect true differences.

Conclusions:

• Measurement error is a significant source of variability affecting inferences from study samples.
• Concerns about measurement accuracy and precision, including TEM, are universal across scientific disciplines.
• Computer-assisted data collection facilitates the incorporation of repeated measures, improving the chances of identifying true biological differences.