Accuracy and precision are common terms that scientists use to measure the amount of error in their data. While they may sound like synonyms, they are not quite the same- we will break it down in this article!
Topics Covered in Other Articles
What are Accuracy and Precision?
When scientists conduct experiments, they often need to ensure that their data and procedures are robust. To do this, they can measure the accuracy and precision of their data to check the amount of error that is present. But what do these terms even mean? Simply put:
Accuracy is how close a measurement is to the actual correct value. Precision is how close measurements are to each other. Accuracy is typically associated with “correctness”, while precision is typically associated with “reliability”. This difference is best illustrated with the classic dartboard example:
Example of Accuracy vs. Precision
- Low accuracy, low precision: in this case, the darts are very far away from each other, demonstrating low precision. They are also extremely far off from the bulls eye, so this is low in accuracy as well.
- High accuracy, low precision: in this case, the darts are far away from each other, demonstrating low precision. However, the darts are fairly close to the bulls eye, so this would be high in accuracy.
- Low accuracy, high precision: in this case, the darts are very close to each other, demonstrating high precision. However, they are also extremely far off from the bulls eye, so this is low in accuracy.
- High accuracy, high precision: in this case, the darts are very close to each other, demonstrating high precision. They are also right on target at the bulls eye, so they are high in accuracy.
Accuracy and Precision: Related Errors
There are various errors that can lead to fluctuations in accuracy and precision. There are two main types of error. One is a systemic error, an error that is consistent in all measurements. This can lead to low accuracy, since it can sway the measurement away from the correct value, but does not impact precision much, since every measurement is moved by the same amount. In our archery example, an example of a systemic error would be if the archer’s bow was warped left. The arrows would consistently miss the target and fly too far left, but they should still land relatively close to each other.
The other type of error is a random error, which varies at random. These are random fluctuations that are not consistent between measurements. This can lead to low precision, since measurements may vary from each other. However, this does not impact accuracy much. Since the errors are random, they should average to 0 across a large data set. In our archery example, an example of a systemic error would be if the wind (assuming the wind blows randomly in various directions) impacts the arrows. The arrows would not land as close together, depending on the direction of the wind, but should still land relatively close to the target.
Youtube Video Lecture on Accuracy and Precision
Please enjoy our animated lecture video further explaining accuracy and precision.
- Classify the following as a systemic error, or a random error.
- A scientist attempts to measure the height of various children, but their measuring device is installed two inches higher than it should be.
- Various bakers attempt to use the same recipe, but differences in humidity lead to different amounts of flour being used.
- A mother uses a probe to measure the temperature of her children when they are sick, but the probe has been calibrated incorrectly.
- A scientist is measuring the mass of various Hershey’s kisses. They obtain values of 5.5, 5.6, and 5.7 grams. The true value of a Hershey’s kiss is 4.5 grams. Is the scientist’s data accurate? Is the scientist’s data precise?
- A scientist is trying to measure the time it takes for a model car to travel down a hill. However, due to their poor eyesight, they start the timer once the car has already started moving for one second. Does this impact the precision or the accuracy? How so?
Practice Problem Answers
- systemic error, random error, systemic error
- The scientist’s data is precise, but not accurate.
- This impacts the accuracy. The time will be consistently short one second.