Reliability Of Scales
Scales are used to judge the specific attributes of a group or object and are designed to return a specific measurement or marker that can be used for analysis. Sometimes this analysis is simple, such as weight and sometimes scales are used for more complex or theoretical measurements. Regardless of the medium, scales are not perfectly accurate and different methods are used to improve them.
Other People Are Reading
-
Variance
-
Depending on the scale, each scaling device will give a variety of different answers. There will always be a difference between the scale's variance of answers and the true variance of the measurements. These variances, and the scale's error, can be analyzed statistically. Regardless of the scale, the same kind of formulas can be used to judge scale reliability.
Cronbach's Alpha
-
The most common "index of reliability" used when analyzing the reliability of scales is known as Cronbach's Alpha. This is a measurement of internal consistency and can be expressed as a function of the number of total test items, along with the average inter-correlation between them. This produces an alpha coefficient ranging from zero to one. The closer to one, the more accurate the scale is.
-
Types of Scales
-
In addition to surveys and other types of statistical tools where variance is important, various mechanical and digital devices designed to return constant responses are also measured using statistical processes. Mechanical scales tend to vary based on the physical properties of the materials they are made from, the springs and dials that must be constantly adjusted in order to give proper results. In digital scales, the electrical signals tend to more accurate but still suffer from slight variance due to minor factors, including heat and moisture.
Emotional Response
-
In surveys, the ordinal indexes used as scales compare responses or points of data. When humans are involved, emotional responses to surveys and polls will always affect variance. A truly accurate human response, from the point of view of the analyzers, is impossible. Since there will always be variance due to unique human responses (people not understanding the question or having previous biases), scale reliability tests can be conducted to help produce more accurate polls.
Reliability of Data
-
Data is not always conducive to examination. Mechanical and digital scales typically operate within the boundary of a specific tolerance for error that has already been tested by the manufacturers. Reliability tests are used only in the first stages of product creation. They are used more frequently when examining scales in relation to statistical data returned from specific questionnaires.
-
References
- Photo Credit scales image by Aleksandr Ugorenkov from Fotolia.com
