1.4
Correlation and Causation
Results of an experiment may suggest that the independent and dependent variables are related.
The relationship between variables, the correlation, can be positive, both variables increase or decrease together. Or negative, one increases and the other decreases.
Additionally, there may be no relationship between the variables. To determine if an apparent correlation reflects a direct cause-and-effect association, a causal relationship, additional control experiments must be performed.
For example, consider an ecosystem where geckos, parasitic ticks, and crows coexist. Crows prey on geckos, and ticks feed on animal blood.
A researcher examines five different gecko populations and finds that the number of geckos without tails decreases as the number of parasitic ticks increases, showing a negative correlation.
Based on this negative correlation alone, the researcher cannot tell whether the parasite directly causes tail loss.
However, if the researcher had counted the number of crows at each location, he may have found a positive correlation between the number of crows and the number of tailless geckos.
And after examining the crows' stomach contents, he would have also found the missing gecko tails.
Together, these observations suggest that crows cause tail loss in geckos.
But a correlation between two variables does not necessarily establish causation. A third variable may affect both variables, creating a correlation between them.
Correlation and Causation
Statistical tests can calculate whether there is a relationship, or correlation, between independent and dependent variables. A relationship between variables shows correlation, but it does not show cause-and-effect. A direct cause-and-effect relationship requires additional controlled experiments. If no consistent relationship exists between the variables, then there is no correlation.
Correlation versus Causation
If the dependent variable increases or decreases when the independent variable increases, there is a positive or negative correlation, respectively, between the two variables. A correlation shows that variables change together, but it does not mean that one variable causes the other.
For example, if a researcher wants to find the cause of tail loss in five different gecko populations and finds a negative relationship between the number of geckos without tails and the number of parasitic ticks, this result shows a negative correlation and suggests that the ticks may not be directly causing tail loss in geckos.
However, if the number of crows near each gecko population is counted, a positive relationship between the number of crows and tailless geckos may be found. If, after examining the contents of the crows' stomachs, the missing gecko tails are discovered, the number of crows would directly affect the number of tails lost by the geckos—showing causation. The earlier correlation could have been coincidental if the gecko tails had not been found in the crows' stomachs.
Importantly, a positive or negative correlation does not mean that causation is present or absent. For example, time spent running and body fat show a negative correlation, and running can directly reduce body fat. On the other hand, ice cream sales and shark attacks show a strong positive correlation, but warmer weather increases both, not one another.
In this example, there may also be a relationship between the number of crows and the number of parasitic ticks. The number of ticks may increase as the number of crows increases, showing a positive correlation. This may also lead to a positive correlation between ticks and tailless geckos. However, unlike the relationship between crows and tailless geckos, ticks and tail loss are not directly connected by cause and effect.
Results of an experiment may suggest that the independent and dependent variables are related.
The relationship between variables, the correlation, can be positive, both variables increase or decrease together. Or negative, one increases and the other decreases.
Additionally, there may be no relationship between the variables. To determine if an apparent correlation reflects a direct cause-and-effect association, a causal relationship, additional control experiments must be performed.
For example, consider an ecosystem where geckos, parasitic ticks, and crows coexist. Crows prey on geckos, and ticks feed on animal blood.
A researcher examines five different gecko populations and finds that the number of geckos without tails decreases as the number of parasitic ticks increases, showing a negative correlation.
Based on this negative correlation alone, the researcher cannot tell whether the parasite directly causes tail loss.
However, if the researcher had counted the number of crows at each location, he may have found a positive correlation between the number of crows and the number of tailless geckos.
And after examining the crows' stomach contents, he would have also found the missing gecko tails.
Together, these observations suggest that crows cause tail loss in geckos.
But a correlation between two variables does not necessarily establish causation. A third variable may affect both variables, creating a correlation between them.
From Chapter 1:
Now Playing
1.4 : Correlation and Causation
Scientific Inquiry
37.6K Views
1.1 : The Scientific Method
Scientific Inquiry
226.1K Views
1.2 : Inductive Reasoning
Scientific Inquiry
60.4K Views
1.3 : Deductive Reasoning
Scientific Inquiry
60.8K Views