Measures are the measurable quantities that make up our physical world. Standards for measurement have a long history in human society and have been developed out of necessity and convenience.
For example, without measurement we could not determine if a piece of furniture was the right size for a room. We also could not keep track of our financial incomes and expenses.
Units of Measurement
A unit of measurement is a standardized quantity used as a factor to express occurring quantities of a physical property. For example, the unit of length is a meter. The unit of weight is a kilogram. The unit of volume is a cubic centimeter. Historically, many of these units have been based on the dimensions of the human body, for example the cubit, pace and hand. They are also based on agriculture, such as the furlong and acre.
As science developed, a need arose for a standard system of measurement that would relate different traditional units of measure to each other. This led to the development of the metric system, which is now used in most countries. When referring to measurements, it is best to spell out the unit of measurement when it begins the sentence or paragraph, especially if readers may not recognize the symbol. For instance, write “We sterilized five 50-mL test tubes.” Also write out the unit when it follows a number that starts the sentence, for example two.
Measurement Methods
The measurement methods chapter presents a set of common functions that need to be built to take the measurements and process the results. It describes what raw measurement data needs to be collected, what calculations need to be made with it and what comparisons are required to detect changes and make the results meaningful.
This chapter also describes the various types of measurement scales that can be used, and how they differ from one another. It also discusses the different ways that empirical relations can be mapped onto numerical ones, which determines scale properties.
It also delves into the way that two general lines of philosophical thought have influenced thinking about method effects and measurement in particular. These strands of thought are often not discussed in the research community, but nevertheless have a strong influence on how researchers think about measurement and method errors. In particular, the doctrine of representationalism has a strong influence on thinking about measurement in human sciences.
Measurement Criteria
The measure focuses on a health outcome, process or structure of care that has the potential for substantial impact. The measure focuses on a nathional health priority that is not currently being adequately addressed in healthcare and has the potential to improve patient-oriented outcomes or reduce unnecessary use of resources.
The measures are evidence-based and important to making significant gains in healthcare quality.
There is opportunity for improvement, even though the performance levels of some processes may have reached near 100% and appear to have been “topped out” by the 2010 Evidence Task Force. In the case of a composite performance measure, there are rational justifications for merging or combining related and/or competing measurements to address harmonization (at the conceptual level) and to resolve stewardship issues.
For a composite performance measure, the components must be rationally linked and weighted, avoid all-or-none scoring, and have an appropriate burden for collecting data and implementing improvements. For eMeasures, validation testing of the computed performance score is also needed.
Measurement Errors
Measurement errors are the deviations between a measured value and its true value. Errors can be either random or systematic. Random errors are fluctuations that can be evaluated statistically (using the standard error of measurement) while systematic errors are inaccuracies that are reproducibly in the same direction and are difficult to evaluate or correct.
It is important to recognize that no measurement of a physical quantity is ever completely accurate. Errors can be corrected through a variety of procedures, including brainstorming with peers about all the potential environmental and methodological factors that might influence results, double checking data entry by entering it on two different machines, piloting experiments under controlled conditions, and using multiple measures for the same construct to triangulate. This is often called error analysis or uncertainty analysis. This is different than validation, which refers to whether the instrument actually measures what it claims to measure. It is also different than reliability, which relates to the reproducibility of a measure.