I was talking to one of my clients, and he said, “Hiren, I have seen many of my friends are so expert in Statistical Analysis that by looking at data, they can say which hypothesis test can be used. I want to learn that”.

CEO and Founder of Stat Modeller

I just smiled and replied that is not a big deal to decide which test could be used by looking at the data if we know the right trick.

I usually tell in my all training programs that for music, the base is “Sa Re Ga Ma Pa Dha Ni Sa”. If you learn that, you can play music irrespective of the language of the song. Similarly, one can easily find the analysis to be performed on the data by just knowing its Data Types.

Data Type is a very generic word as has many definitions concerning different segments. If you ask developers, they may say, eNum, object, string, boolean, etc… If you ask ML engineers, they may say, float, integer, string, logical, time series etc… If you ask Six Sigma (operational Excellence) professionals, they would say, it's variable and attribute or sometimes they say discrete and continuous. When you ask any researchers they will say, Nominal, Ordinal, Interval and Ratio.

Here, I am going to discuss with respect to the researcher as it is in detail as well as useful for the statistical analysis decisions.

Mainly data can be categorized into two categories such as quantitative and qualitative.

**Qualitative Data — **Data which can not be measured. It can only be counted. Such as sex, types of injury, type of defects, machines etc…

Qualitative data further can be categorized into Nominal and Ordinal Data.

**Nominal Data: **It is unordered categorical data as order or the categories are not important. They give us insight that our object belongs to some category. Even a serial no. or roll number are also a nominal type of data.

**Example #1 **Sex — there can be two categories, male or female.

**Example #2 Color —** If I ask 100 people What is your favorite color, then it could be black, red, green, etc…

**Ordinal Data:** It is ordered categorical data as the order of categories makes some sense and is important. Rank is one of them.

**Example #1** Type of Injury — Near Miss, Minor, Major, Fatal. These are 4 categories of injury and they are ordered with respect to their severity. This also can be organized in ascending or descending order.

**Example #2 **Socio-Economic Class — Low, Medium, High.

Here, in ordinal data, we can know the order from ascending to descending but we can not know how much difference is there. e.g. There are Top 5 students with their rank. By just knowing their ranks, we can not get the idea that Ranker 1 is how much ahead of others and Ranker 5 is how much behind others.

**Quantitative Data — **Data which can be measured. I usually say that for any data collection, you need any measuring instrument is a quantitative data.. (However, it is just to understand. Not 100% true. e.g. My engineer friends would know that Go-No-Go Gage which is an instrument but it generates only qualitative data) Such as length, width, height, diameter, age etc…

Quantitative data further can be categorized into Interval and Ratio Data.

**Interval Data: **Interval data not only gives us an insight into the difference between two values. Such as temperature. Interval and Ratio data look very similar. The key difference between them is Zero. What, Zero? Yes, Zero.

In Interval data, the value of zero is not true zero, it is arbitrary. e.g. Zero Temperature doesn’t mean that temperature does not exist. It also can be negative.

**Ratio Data: **This type of data looks very similar to the interval data. You need to have some measuring instrument to record the data.

Unlike interval data, here value of zero is true which means it is absolute. e.g. length is zero which means no length exists. It can’t be negative.

**Examples: **height, weight, length, diameter, age, BMI (Body Mass Index) etc…

The selection of statistical analysis tools and techniques depends upon the data types. The image below summarizes it well.

Hope this article will help you to choose the right tools and techniques for your data analysis. If you want to learn about how to perform analysis on your research data, you can visit this course on Make your Research Effective using SPSS.

Happy Learning!!!

*Originally published May 2, 2019*