Causal Research: Identifying Relationships and Making Business Decisions through Experimentation

We are at the final stop on our crash course on the three types of survey research. Over the last month we have gone over both exploratory and descriptive research. Today we will finish off our blog series by jumping into the world of causal research. This article will take us through the purpose of causal research, how to implement it in your research projects, and some great examples of how organizations are currently using causal research to make better business decisions.

What is Causal Research, and Why is it Important?

Causal research falls under the category of conclusive research, because of its attempt to reveal a cause and effect relationship between two variables. Like descriptive research, this form of research attempts to prove an idea put forward by an individual or organization. However, it significantly differs on both its methods and its purpose. Where descriptive research is broad in scope, attempting to better define any opinion, attitude, or behaviour held by a particular group, causal research will have only two objectives:

  1. Understanding which variables are the cause, and which variables are the effect. For example, let’s say a city council wanted to reduce car accidents on their streets. They might find through preliminary descriptive and exploratory research that both accidents and road rage have been steadily increasing over the past 5 years. Instead of automatically assuming that road rage is the cause of these accidents, it would be important to measure whether the opposite could be true. Maybe road rage increases in light of more accidents due to lane closures and increased traffic. It could also be the case of the old adage ‘correlation does not guarantee causation.’ Maybe both are increasing due to another reason like construction, lack of proper traffic controls, or an influx of new drivers.
  2. Determining the nature of the relationship between the causal variables and the effect predicted. Continuing with our example, let’s say the city council proved that road rage had an increasing effect on the number of car accidents in the area. The causal research could be used for two things. First measuring the significance of the effect, like quantifying the percentage increase in accidents that can be contributed by road rage. Second, observing how the relationship between the variables works (ie: enraged drivers are prone to accelerating dangerously or taking more risks, resulting in more accidents).

These objectives are what makes causal research more scientific than its exploratory and descriptive counter parts. In order to meet these objectives, causal researchers have to isolate the particular variable they believe is responsible for something taking place, and measure its true significance. With this information, an organization can confidently decide whether it is worth the resources to use a variable, like adding better traffic signs, or attempt to eliminate a variable, like road rage.

Implementing Causal Research Effectively

Causal research should be looked at as experimental research. Remember, the goal of this research is to prove a cause and effect relationship. With this in mind, it becomes very important to have strictly planned parameters and objectives. Without a complete understanding of your research plan and what you are trying to prove, your findings can become unreliable and have high amounts of researcher bias. Try using exploratory research or descriptive research as a tool to base your research plan on.

Once your research plan and objectives are fleshed out, it’s time to set up your causal experiment properly. Here are three major conditions about your causal experiment you’ll want to check off before you set it into motion:

  1. The cause and effect relationship will be proved or disproved by the experiment. Of course this may seem like a no-brainer, but if you do not make sure your research plan directly ties into your research objective, the end results of your study will be as fruitless as most children’s cereals (no offense Tucan Sam). To make sure your study will have results one way or another, observe what your normal environment is and then crank up the frequency or power of the causal variable.
  2. You are clearly identifying which variables are being tested as independent (causing effect) and which are being tested as dependent (being effected). As discussed in the road rage/car accident example, in many cases it is hard to tell which variable is dependent on the other. Because of this, it is essential to identify which will be tested as which prior to the experiment. Usually, the independent variable will be whatever you are adding to the environment.
    For example, we hypothesize that increasing colour options for cars will increase sales. In this case, the number of colour options is the independent variable and the level of sales is our dependent variable. Your next step would be to measure the normal rate of sales at the car store, and then add a broader selection of car colours. After collecting the new sales numbers, compare the two data sets and study the effect on sales.
  3. There are no external variables that can also be causing changes in your results. Without accounting for all possible factors that might effect changes in your dependent variable, you can’t be certain it is the variable being tested that is truly responsible for causing the effects we measure. In the laboratory, scientists have the luxury of being able to create a completely neutral environment. Unfortunately for the rest of us, we have to deal with the environment we are given. So the most important thing to do when creating your research plan is to ensure that your experiment occurs under the most similar possible conditions as when you measured your normal results.
    For example, let’s say you are an ice cream store owner and want to study the effect a clown handing out balloons in front of your store will have on sales. Awesome idea, I know! It would be a bad idea to use your summertime sales as your normal data source and run your experiment in winter. Not only would that be cold for the clown, the weather would have a huge effect on ice cream sales.

Who Uses Causal Research and How Can I Incorporate it in my Business Goals?

It really doesn’t matter what type of organization you are or what goals you have, causal research can be used to benefit you. The goal of causal research is to give proof that a particular relationship exists. From a company standpoint, if you want to verify that a strategy will work or be confident when identifying sources of an issue, causal research is the way to go. Let’s take a look at a few examples of how causal research could be implemented with different goals in mind:

  1. Increasing Customer Retention: Most franchise chains conduct causal research experiments within their stores. In one case, a large auto-repair shop recently conducted an experiment where select shops enforced a policy that an employee would have a one-on-one with the client while their vehicle is being assessed. They were instructed to go over any concerns and speak in layman’s terms about anything wrong with the car, focusing on the client understanding the issues.
    This experiment was implemented because of an online survey that identified a lack of employee-client communication as being a barrier to repeat customers. After identifying two solutions to this issue (facilitating discussion and increasing client understanding), the company used this experiment to learn just how effective these solutions would be in increasing customer retention. By comparing the sales in unchanged shops to those that were part of the experiment, the company noticed a significant increase in customer loyalty.
  2. Community Initiatives: City councils often use causal research to measure the success of their community initiatives. Let’s say the City of Ottawa conducted a survey and learned that Ottawan’s were dissatisfied with current public transit options. They could then set in motion a strategy to create more ‘Park and Rides’ to help more people be able to ride the bus. After implementing this strategy they can resend the same survey and measure what type of effect it has had on the overall satisfaction of public transit.
  3. Effective Advertising: Advertising is one of the most common sectors for causal research. Most times companies will test ad campaigns in small areas before expanding it across all locations. The idea is to measure whether there is a sufficient increase in sales, leads or public interest in those regions with the advertisement before committing fully.
    Many organizations will take this experiment a step further by creating a survey asking customers what made them visit or interested in their services. Now the business can compare responses from customers in the experiment area to the responses of their overall client base and see if there increase in traffic is a direct result of their advertising.

Get Your Research On!

Congratulations, we have just completed our four part survey research crash course! With your newfound knowledge of exploratory, descriptive, and causal research, you’ll be able to create the perfect research plan to take advantage of any business opportunity. The next step is to learn how to avoid error and bias in your research, so join me on another FluidSurveys University mini-series by visiting part one, “How to Know the Difference Between Error and Bias .”

FluidSurveys Presents

Free Survey Q&A

Join our survey & research expert Rick Penwarden as he answers all of your questions every Wednesday at 1PM EST!


Leave a Reply

Your email address will not be published. Required fields are marked *