Discrete Data

Discrete data is data that is countable, quantitative data. It is numerical in data, stemming from a physical count or measurement, and has a limited number of possible values. The number of wheels on a vehicle would be discrete data.

Discrete data has the advantage of being easy to collect. It also has the drawback of being limited in its use.

Discrete data can be collected by direct observation or can be truncated continuous data. Continuous data, as the name suggests, can be any value. Consider the height of students in a classroom. You may use a very accurate tape measure and calculate heights to the thousandth of an inch, with any value that you come up with being valid.

You may not need that level of detail for your analysis, though, and may decide instead to go with discrete data. You could use the number of inches, or even double that countable number by giving half inch increments as an option. That would limit you to options 48, 49, 50, 51, on up to maybe 78 inches or so as the possible values. You would double that number if you added in half inch increments.

But the point is that after you collect the data, you have countable buckets. You can still ‘math up’ the data and rank, average, and slice & dice however you want.

The challenge, though, comes if you need very precise information to come to the conclusions you are looking for. As you could imagine, if you had precise heights of every student, you could get a lot more detailed in your analysis.

Keep in mind that continuous data can be turned into discrete data, but the reverse is not true. You can’t turn discrete data into continuous data.

Why not always collect continuous date? Well, first off, sometimes discrete data is the only option. The number of eggs laid per chicken per day can’t be continuous data, though you could use total weight of the eggs as continuous data.

The other reason is that continuous data takes more effort to collect. Looking in the nest and pointing at each egg as you count is simple. Collecting, weighing, and recording the weight of the eggs is a lot more time consuming of a process.

Some people use this term interchangeably with ‘attribute data’.

While the configuration of the data is similar, a countable number, the basis of the data is not. Attribute data is derived from a qualitative characteristic. This may be sequenced (ordinal data, as in high, medium, or low), binary (yes/no), or nominal (unsequenced data as in red, yellow blue). Discrete data is quantitative at its heart.

The number of pets per household would be discrete data. If you characterized the pets as ‘furry’, ‘scaly’, or ‘feathery’, you still end up with countable buckets, but it is based on a qualitative assessment, so would be attribute data.

The distinction might be murky if you organized by species instead. You could argue that you are actually collecting several types of discrete data (i.e. the number of snakes is one data collection effort, the number of dogs is another.) You could also look at it as a set of nominal data that you are collection.

Try not to get into the weeds on things like this. Focus on the intent of why you are collecting the data, how you will gather it, and what you are going to do with it. In the end, regardless of what you call it, the use is what matters.