ALL MEASUREMENT IS APPROXIMATE

I wish more people understood the simple fact that all measurement is approximate. Everything we measure has an inherent amount of error, a degree to which the measured value floats around the actual value. Furthermore, people need to understand that this actual value can never be known. Thus, you may tell people that you are 5 feet, 9 inches tall, but this is only an approximation, and furthermore, there is no tool in the entire universe that can determine your actual height! Sure, you can reduce the amount of error in a measurement by getting better tools, and you can use this super-advanced tool to find that you are 5 feet, 9.00342 inches tall, but even with its advanced technology, there will remain a degree of error in that measurement.

Unfortunately, your intelligence cannot be measured in the same way that we measure your height. Your intelligence is a "latent trait" which means it can't be seen or touched, so we create standardized tests to measure this latent trait in the same way that a ruler is used to measure your height. What is problematic is that many people seem to think that this measurement is definitive and static, but nothing could be further from the truth.

People need to realize that these measures of latent abilities (standardized tests) are subject to errors of measurement in the same manner that your ruler is subject to errors in measurement. Now, I am not referring to the usual “errors” that people raise about standardized tests, such as “it was too hard”, “the questions were stupid”, or “the DOE continues to change the tests and we don’t know what is coming.” These are complaints, and while some may be valid, these are not the sources of error to which I am referring.

In a standardized test itself, there are numerous approximations and probability-based assumptions that cause for an inherent amount of error within the test. You could look into the many intricacies and details behind “Item Response Theory” which determines how tests are scored for the Algebra I Core 40 Exam, or you could just observe the simple fact that for standardized tests, you could repeat the exam and receive a significantly different score. Thus, repetitive measures will sometimes produce significantly different results. This is problematic.

Allow me to be clear: I am not complaining about standardized tests. I think they are a useful tool to measure something that is very difficult to measure. However, I think people need to realize the volatility of these tools. Their accuracy in measurement is not perfect because all measurement is approximate. Your standardized test score is not a definitive snapshot of your intelligence; rather, it is a blurry image that can only make sense in the background of a much larger set of data and observations.

If we removed standardized tests because they contained error, then we might as well never measure the miles-per-gallon of a car, the height of a building, or the number of people living in Indiana. If we removed standardized tests because they contained error, then we should just get rid of all measurement in our lives because all measurement contains error.

All measurement is approximate. Look into how standardized tests are scored, and you will realize just how approximate this measure actually is.