# Third Variable Problem: Definition & 10 Examples

Reviewed by Dalia Yashinsky (MA, Phil)

Dalia Yashinsky is a freelance academic writer. She graduated with her Bachelor's (with Honors) from Queen's University in Kingston Ontario in 2015. She then got her Master's Degree in philosophy, also from Queen's University, in 2017.

The “third variable problem” refers to a situation in statistical analysis where an association between two variables might be better explained by a relationship with a third variable, which is not considered in the analysis.

As Bernardo Carducci of Indiana University Southeast defines it,

“The third-variable problem exists when the observed relationship between two variables is actually produced by their relationship with another unobserved, or third, variable.” (Carducci, 2009)

This third variable is known as the confounding variable or lurking variable.

It is considered a problem because, when not observed and controlled, it may interfere with the integrity of the results and lead to the false cause fallacy. As Pozder (2018) notes:

“The failure to control the third variable may bring about a correlation that is false between the independent and dependent variables.”

Let’s look at some examples.

## Third Variable Problem Examples

### 1. The Impact of Ice Cream Sales on Drowning Incidents

Independent Variable: Ice Cream Sales
Dependent Variable: Drowning Incidents
Third Variable: Temperature (Specifically, hot weather)

It might be observed that as ice cream sales increase, the number of drowning incidents also rises. We might prematurely conclude that buying more ice cream has a direct effect on increasing the number of drownings. However, this is where the third variable problem comes into play. The actual third variable influencing both ice cream sales and drowning incidents is hot weather. During hot weather, people are more likely to buy ice cream because it’s refreshing. Simultaneously, they are also more likely to go swimming, which can lead to an increased risk of drowning.

### 2. The Connection Between Firefighter Numbers and Property Damage

Independent Variable: Number of Firefighters at a Scene
Dependent Variable: Amount of Property Damage
Third Variable: Severity of the Fire

When there are more firefighters at a scene, there might be a correlation with higher property damage. One might hastily conclude that having more firefighters present results in more damage. The third variable problem reveals that the actual correlation is with the severity of the fire. Larger, more destructive fires would naturally require more firefighters to combat them and simultaneously result in more property damage.

### 3. The Link Between Nighttime Lighting and Sleep Quality

Independent Variable: Amount of Nighttime Artificial Lighting (e.g., street lights, neon signs)
Dependent Variable: Sleep Quality of Residents
Third Variable: Urban Development and Noise Levels

In areas with significant nighttime artificial lighting, there might be reports of poorer sleep quality among residents. It could be mistakenly believed that the lighting alone is the primary disruptor of sleep. However, the third variable problem brings in the influence of urban development and noise levels. Densely developed urban areas might have more nighttime lighting, but they also have higher noise levels due to traffic, nightlife, and other factors that can equally or more significantly disrupt sleep.

### 4. The Correlation Between Park Count and Happiness Levels

Independent Variable: Number of Parks in a City
Dependent Variable: Reported Happiness Levels of Residents
Third Variable: Overall City Infrastructure and Planning

Cities with a higher number of parks might demonstrate increased reported happiness levels among its residents. One might mistakenly believe that just the presence of parks directly boosts happiness. The third variable problem highlights that the overall infrastructure and planning of the city can influence both. Well-planned cities might prioritize parks and green spaces, but they might also have efficient public transport, less traffic congestion, and better amenities, all contributing to the happiness of its residents.

### 5. The Association Between School Libraries and Student Performance

Independent Variable: Number of Books in School Libraries
Dependent Variable: Academic Performance of Students
Third Variable: Overall School Funding and Resources

Schools with vast libraries and a high number of books might show students performing better academically. It could be mistakenly believed that merely the size of the library boosts student performance. The third variable problem highlights the influence of the overall school funding and resources. Schools with ample resources might have well-stocked libraries, but they also likely have better facilities, more qualified teachers, and superior learning environments that collectively contribute to improved student performance.

### 6. CPU Speed and Computer Longevity

Independent Variable: Speed of a Computer’s CPU
Dependent Variable: Longevity of the Computer
Third Variable: Quality of Other Hardware Components

Computers with faster CPUs might appear to last longer. One might quickly assume that a faster CPU directly ensures the computer’s longer lifespan. However, the third variable problem highlights the overall quality of other hardware components. Computers equipped with high-speed CPUs are often premium models, which also come with high-quality motherboards, power supplies, and cooling systems. It’s the collective quality of all these components that can contribute to the computer’s longevity.

### 7. Brightness of a Star and Its Size

Independent Variable: Brightness of a Star
Dependent Variable: Size of the Star
Third Variable: Distance of the Star from Earth

In astronomical observations, brighter stars might appear larger to telescopes. One could think that the brighter a star is, the larger its size. However, the third variable problem introduces the factor of the star’s distance from Earth. Stars that are closer to Earth can appear brighter and might be misinterpreted as being larger when, in reality, their apparent brightness is due to their proximity to our planet.

### 8. The Link Between Height and Self-Esteem

Independent Variable: Height of an Individual
Dependent Variable: Self-Esteem Level
Third Variable: Attractiveness as Perceived by Society

A study claims a positive correlation between height and self-esteem. One might initially think that taller height directly leads to higher self-esteem. However, the third variable problem points out the influence of attractiveness. In many cultures or societies, taller individuals might be perceived as more attractive, and this perceived attractiveness can boost their self-esteem, independent of their height.

### 9. Job Satisfaction Among Older Employees

Independent Variable: Age of the Employee
Dependent Variable: Job Satisfaction Level
Third Variable: Job Position Held in the Company

A study suggests that older employees tend to be more satisfied with their jobs. At first glance, it might seem that age directly correlates with job satisfaction. However, the third variable problem introduces the influence of the job position. Older employees, due to their tenure and experience, might hold higher, more rewarding positions within the company, leading to increased satisfaction, independent of their age.

### 10. Seasonal Depression and Light Exposure

Independent Variable: Reduced Light Exposure in Winter
Dependent Variable: Rates of Depression
Third Variable: Level of Physical Activity

Research into seasonal depression might indicate a direct relationship between reduced light exposure during winter months and increased rates of depression. However, the third variable problem sheds light on another significant factor: physical activity. People often engage in less physical activity during the colder, darker winter months. This reduced activity can also influence mood and contribute to depression, making it a potential confounder in the relationship between light exposure and depression.

## Common Causes of the Third Variable Problem

There are countless causes for the third variable problem. However, below are some common factors that need to be taken into account by skilled researchers to control for third variables.

### 1. Incomplete Understanding of Underlying Mechanisms

One of the primary causes of the third variable problem is the researcher’s incomplete or insufficient understanding of the underlying mechanisms governing the phenomena under investigation (Boniface, 2019).

If researchers don’t fully comprehend all the intricate processes, interactions, or factors influencing their variables of interest, they might overlook potential confounders.

For instance, a study might examine the relationship between dietary habits and health outcomes without considering the role of physical activity, leading to erroneous conclusions.

### 2. Over-reliance on Observational Data

While observational studies play a crucial role in research, they’re particularly susceptible to the third variable problem.

As Johnson and Christensen (2019) note:

The most serious problem that we run into in the simple cases of nonexperimental research is that the observed relationship might be due to an extraneous variable, and this problem is widespread in nonexperimental research.

Unlike experimental designs, where researchers can control and manipulate variables, observational studies simply record data as it appears in the natural environment.

This passive observation means that hidden variables, not immediately apparent to the researchers, can influence the observed association.

For instance, observational studies might show a link between coffee consumption and longevity, but if they don’t account for factors like socioeconomic status or overall lifestyle, the findings could be misleading.

See More: The 3 Types of Experimental Design

### 3. Insufficient Data Collection

At times, the third variable problem arises from the sheer lack of comprehensive data.

If researchers do not collect detailed, varied, and extensive data, they might miss out on potential confounding variables. A study might focus on specific factors while neglecting others due to resource limitations, biases, or oversight (Parker & Berman, 2016).

For example, a survey exploring the relationship between screen time and sleep quality might neglect to inquire about caffeine consumption, a potential confounder affecting both variables.

### 4. Researcher Bias

Every researcher brings a set of biases—conscious or unconscious—to their work.

These biases can shape the design, implementation, and interpretation of a study. If a researcher holds strong beliefs or hopes about a particular outcome, they might unintentionally design their study in a way that overlooks or minimizes potential third variables (Nestor & Schutt, 2018).

This selective attention or confirmation bias can lead to an overemphasis on the primary relationship of interest while ignoring confounding variables.

## References

Boniface, D. R. (2019). Experiment Design and Statistical Methods For Behavioural and Social Research. CRC Press. ISBN: 9781351449298.

Carducci, B. J. (2009). The Psychology of Personality: Viewpoints, Research, and Applications. Wiley.

Johnson, R. B., & Christensen, L. (2019). Educational Research: Quantitative, Qualitative, and Mixed Approaches. SAGE Publications.

Nestor, P. G., & Schutt, R. K. (2018). Research Methods in Psychology: Investigating Human Behavior. SAGE Publications.

Parker, R. A., & Berman, N. G. (2016). Planning Clinical Research. Cambridge University Press.

Pozder, D. (2018). Without Stigma: About the Stigma and the Identity of the Mental Illness. Xlibris AU.

Dr. Chris Drew is the founder of the Helpful Professor. He holds a PhD in education and has published over 20 articles in scholarly journals. He is the former editor of the Journal of Learning Development in Higher Education. [Image Descriptor: Photo of Chris]

Dalia Yashinsky is a freelance academic writer. She graduated with her Bachelor's (with Honors) from Queen's University in Kingston Ontario in 2015. She then got her Master's Degree in philosophy, also from Queen's University, in 2017.