Identifying and Validating Scales

Some researchers use existing scales that fit their construct and domain, some modify a published scale for a new study, and others develop their own. When utilizing an existing scale, it is important to identify the type of scale used (see Table 3), validate the scale with appropriate procedures (see Table 5), and report exactly how it was deployed.


In math, scale is a ratio that compares corresponding sides of two figures. This concept is also applied to models, maps, blueprints and the scale you weigh yourself on. Ratios are all around us, but some are more obvious than others, such as the scale on a map or blueprint or the ratio of ingredients when making cement.

Scale is also the set of tones that forms a musical mode. The most common scale is the diatonic scale, but there are many different scales used in music throughout the world. Each scale has a characteristic interval pattern and a specific starting point (or tonic) note.

The term “scale” can also refer to any thin, platelike piece or lamina that peels away from a surface, as from the skin. It can also mean the flat, horny plates that form the covering of some animals, as snakes or lizards. Finally, it can be used to describe a system of compensation, such as pay scales for actors or musicians.


Historically, scales have been developed in order to improve the accuracy of measurement. They were largely developed through trial and error, with inventors such as Leonardo da Vinci lending their hand to the development.

Scales also play an important role in the analysis of music from nonliterate cultures and folk music. However, their function as theoretical concepts is more prominent in the music of highly sophisticated cultures (variously described as art music, classical music, cultivated music or high culture music).

Musical scales are often taught to students as part of their formal instruction. They may be learned intuitively through experience or taught explicitly using written music theory. Some scales can be identified by their constituent intervals, such as being hemitonic or cohemitonic. Alternatively, they can be recognized by the repetition of characteristic melodic motives, such as the tumbling strains described by Curt Sachs in the singing of Australian Aboriginal peoples. This can help distinguish different types of scales even when they are sung at the same pitch level.


There are four different types of scale: nominal, ordinal, interval and ratio. These classifications describe the level of information recorded within a variable and influence what kind of statistical analyses you can perform on your data.

A nominal scale has categories that you can name and doesn’t have a natural order, such as gender, college major or blood type. It’s the simplest form of measurement and can be used to categorize or label observations. You can either leave these labels as they are or you can code them to identify the groups you want to compare.

An interval scale is a step up from nominal. It allows you to rank your observations in an ordered way and also lets you add or subtract them. You can think of intervals like the temperature in Celsius and Fahrenheit, credit scores and SAT test results. It’s also possible to use ratios on an interval scale, such as when rating someone’s response time, like Amar took 2.3 seconds longer than Becky did.


Scalable apps provide a high-quality user experience and prevent performance issues that can degrade brand trust. They also ensure that applications can accommodate growth without sacrificing performance or adding complexity.

A scale drawing can make it easier to interpret complex objects and structures, such as blueprints or machinery. It can also help architects, machine-makers and designers work with models of objects that are too large to hold. A map scale shows the relative size of geographic features, such as mountains and rivers, by using a ratio. Many maps include both verbal and representative fraction (RF) scales.

There are several ways to make an application more scalable, such as adding more CPUs or increasing memory limits. However, these methods increase the overall speed of the application but don’t address problems that arise from complex interactions between different parts of the software. A better approach is horizontal scaling, which involves distributing workloads across multiple machines in the same cluster.

Posted in News.