1.5: "Quantifying" a Construct, or Why Numbers Are Better Than Words

In a quantitative study, we are trying to quantify a construct.

“Quantifying” means we are trying to use numbers to describe how a construct varies. When something varies, we mean that we see two people or things possessing two different levels of something. Person A is more depressed than person B. This type of abuse is different than that type of abuse. Variation implies a comparison between at least two entities.

Why Numbers are Better than Words

We could use words to describe how something varies on a continuum. This person is more depressed than that person. This abuse is physical, while that abuse is emotional.

But words can shift in meaning depending on the person – "What do you mean by that?”; "I've seen this self-harm before, it's not that big of a deal vs. No, this self-harm is pretty severe"; "Your impression of racism is more severe than my impression."

Instead of using words, numbers help to communicate the variation. With numbers, we can do so much more. We can communicate by how much more, how many more, how much more frequent, and how much more intense. Instead of simply saying “it” is more than “that,” we can quantify the amount of the “more” – “This person is how much more depressed than that person?”; “This person is how much more isolated than that person?”

Using Numbers to be on the Same Page

Communication involves standardization and validity.

Using a standard set of numbers, we are “on the same page,” i.e., we agree on what the numbers mean, not just if two things/people are different, but by how much.

This is called standardization – we agree that these numbers tell us what we want to know.

The number represents the variation as we experience it. That number has to represent one entity's level of variation in comparison to another entity's level of variation. We observe a variation, and that number represents that variation, and we agree that the number represents that variation.

Validity means that these numbers are helpful. It means that we agree that these numbers help us predict something else.

Instead of using biased discussion, we use numbers to (as best we can) have a common discussion framework. One person from Chicago says it is a hot day; the person from the South says it is not that hot. The purpose of a number is to represent the level of “it,” in this case, temperature. Hence, we are on the same page about the level of “it.” Then, we can say separately that the experience of how comfortable the temperature is is independent and varies depending on the geographical region of the person. So, if we say the day’s temperature is 85 degrees, we can agree that number represents the temperature. Our experience of the comfort level of the temperature varies according to a separate “it,” which is the region of the country for each person.

The previous paragraphs use numbers to represent how “it” varies by level, either by low or high amount or low or high intensity. “It” also varies by type. We also use numbers to represent variation by type, but the numbers are simply codes. Numbers as codes do not carry any inherent value of low to high. Gender (again assuming binary) is an “it” that varies by type: male or female. Females receive a code of “1,” and males receive a code of “2.” There is no inherent value of more or less. However, statistics processors cannot use words to compute their statistics, so we must use a number to represent the type. So, when you see numbers representing type, it is just code and nothing more. We will return to this discussion when we discuss variable types.