# Constructing a conceptual framework

A conceptual framework illustrates what you expect to find through your research. It defines the relevant variables for your study and maps out how they might relate to each other.

You should construct a conceptual framework before you begin collecting data. It is often represented in a visual format.

This article explains how to construct a conceptual framework for an expected cause-and-effect relationship, incorporating relevant variables that might influence that relationship.

## What is a conceptual framework?

A conceptual framework is a written or visual representation of an expected relationship between variables. Variables are simply the characteristics or properties that you want to study.

The conceptual framework is generally developed based on a literature review of existing studies and theories about the topic.

A conceptual framework can be designed in many different ways. The form yours takes will depend on what kinds of relationships you expect to find.

## Independent and dependent variables

If we want to test a cause-and-effect relationship, we need to identify at least two key variables: the independent variable and the dependent variable. In our example:

- the expected cause, “hours of study,” is the independent variable (aka the predictor or explanatory variable).
- the expected effect, “exam score,” is the dependent variable (aka the response or outcome variable).

In other words, “exam score” *depends on* “hours of study.” Our hypothesis is that the more hours a student studies, the better they will do on the exam.

Causal relationships often involve several independent variables that affect the dependent variable. However, to keep things simple, we’ll work with just one independent variable, namely “hours of study.”

To visualize our expected cause-and-effect relationship, we will use the basic design components of boxes and arrows. Each variable appears in a box. To indicate a causal relationship, each arrow should start from the independent variable (the cause) and point to the dependent variable (the effect).

Next, we should identify other variables that might influence the relationship between our independent and dependent variables. Some common variables to include are moderators, mediators, and control variables.

## Moderating variables

Now we’ll expand the framework by adding a moderating variable (aka a moderator). A moderator alters the effect that an independent variable has on a dependent variable.

The moderator thus changes the effect component of the cause-and-effect relationship. This moderation is also referred to as the interaction effect.

In our example, we expect that the number of hours a student studies is related to their exam score: the more you prepare, the higher your score will be.

Now we add the moderator “IQ.” A student’s IQ level changes the effect that the variable “hours of study” has on the exam score: the higher your IQ, the fewer hours of study you must put in to do well on the exam.

In other words, the “IQ” moderator *moderates* the effect that the number of study hours has on the exam score.

Let’s take a look at how this might work. The graph shows how the number of hours spent studying affects exam score. The more hours you study, the better your results. A student who studies for 20 hours will get a perfect score.

But the graph looks different when we add an “IQ” moderator of 120. A student with this IQ will already achieve a perfect score after just 15 hours of study.

Below, the value of the “IQ” moderator has been increased to 150. A student with this IQ will only need to invest five hours of studying in order to get a perfect score.

The higher the IQ, the fewer hours a student needs to study in order to achieve a score of 100%.

In short, a moderating variable is something that changes the cause-and-effect relationship between two variables as its value increases or decreases.

## Mediating variables

Now we’ll expand the framework by adding a mediating variable. In a cause-and-effect relationship, a mediating variable is a variable that links the independent and dependent variables, allowing the relationship between them to be better explained.

Here’s how the conceptual framework might look if a mediator variable were involved:

The mediating variable of “number of practice problems completed” comes between the independent and dependent variables. The hours of study impacts the number of practice problems, which in turn impacts the exam score.

In this case, the mediator helps explain *why* studying more hours leads to a higher exam score. The more hours a student studies, the more practice problems they will complete; the more practice problems completed, the higher the student’s exam score will be.

By adding the mediating variable of “number of practice problems completed,” we help explain the cause-and-effect relationship between the two main variables.

Keep in mind that mediating variables can be difficult to interpret, and care must be taken when conclusions are drawn from them.

## Moderator vs mediator

It’s important not to confuse a moderators and mediators. To remember the difference, you can think of them in relation to the independent variable.

A mediating variable *is* affected by the independent variable, and it affects the dependent variable. Therefore, it links the two variables and helps explain the relationship between them.

A moderating variable is *not* affected by the independent variable, even though affects the dependent variable. For example, no matter how many hours you study (the independent variable), your IQ will not get higher.

## Control variables

To test a cause-and-effect relationship, we also need to consider other variables that we’re not interested in measuring the effects of, but that could potentially impact students’ exam scores.

These are control variables—variables that are held constant so that they don’t interfere with the results.

For example, it is likely that if a student feels ill, they will get a lower score on the exam. Therefore, we’ll add “health” as a control variable.

That means we should keep the variable “health” constant in our study—we’ll only include participants who are in good health on the day of the exam.