Data Types & Measurement Scales

Data

Data is a sophisticated & factual base behind any information, without which no theory or prediction can hold any significant value.

Example

  •  Global warming is increasing per year.  (Vague information)
  • Global warming is increasing at 0.25% rate per year.  (Factual thing)

Thus, data is a base of everything, as Albert Einstein beautifully crafted it as,

“   Our lives revolve around the data”

 

Data in Improvement Projects

  “What we can’t measure, we can’t improve”

In everyday industrial manufacturing or service environment, data is being generated. Every system, process and sub-processes are interlined with each other and generating data whether on terms of check sheets, ERP etc. The resultant data is than being scrutinized in the light of statistical software. Now, we will cast a look over adapt toes and its measurement scale.

Data is basically divided into two types. Below is a figure illustrating the phenomenon?

Data TypesOne caution is to always try to shift from attribute data to continuous one due to following reasons,

  • Special Cause Variation is more clearly depicted in Control Charts for continuous data rather than attribute ne.
  • Doe can be more proned to the continuous measurement alike Regression Analysis.

Continues entity holds more information than discrete data. (It can tell us about the range or other descriptive statistics in a more reliable manner).

 

Measurement Scales.

There are basically 4 ties of scale and they hold different kind of tests suitable for their applications. They are termed as “NOIR”.

screenshot-746Nominal Scale

It classifies data with no specific order. It is a lowest level of measurement with no natural order and sometimes referred to as categorical or qualitative.

Example Gender, Marriage Relationship Status etc.

 

Ordinal Scale

It refers to a red in a series but precise difference is not finalized.

Examples: It includes relative ranking like; Likert scale (customer satisfaction), socioeconomic status, Pain intensity. (Not necessarily equal intervals)

It can undergo,

  • All Nominal Level tests
  • Median
  • Percentile
  • IQR
  • Correlation coefficient rank order

Interval Scale

It has meaningful difference between sets but arbitrate zero value. All the categories are well orders and equal distanced. It can include negative numbers.

Example

Examination Marks. (Exact difference between numbers is known)

It can undergo,

  • All Ordinal Tests
  • Mean
  • S.D.
  • +ION & -ION (Can’t be multiplied or divided)

Ratio Scale

It is the most priced scale. It has got meaningful difference and equal distance values. In this scale, zero means none.

Most of the time Interval/Ratio tests are combined for the sake of easiness.

It can undergo,

  • All descriptive & inferential tests.
  • Tests for a Measurement Scale Validation

 

The other type of chart can be,

screenshot-747

Conclusion

Always answer three question before getting started with data and its concerned tests.

  1. What is a type of data?
  2. How many samples do you have?
  3. What is the purpose of your hypothesis? (What do you want to prove?)
Open chat
1
Hello !!

How can we help you?