# Grading Your Own Homework In science you have to determine what you want to measure. Say you are building a system to heat a water bath. You want to determine how accurate it is when heating to certain temperatures. In order to determine this, you must measure the temperature. At first you do this by sticking your finger in the water to guess the temperature. You proclaim that the system is perfectly calibrated because it "feels" write. Your partner argues that this is unacceptable! You are "grading your own homework"! You need a means of measuring the water temperature that avoids this. He proposes you use a thermometer. You do this and the bath still looks well calibrated. He now proclaims "see, we are no longer grading our own homework - we have a reliable, independent measurement". But do you? Is it true that you are really no longer grading your own home - i.e. assessing the calibration of the water? No! You will _always_ be grading your own homework. In the case of the thermometer, you still must have a theory of why it is accurate, why it is not making errors, and so on. If you don't have that theory, you could have a rigged thermometer that spits out the answer you are looking for. The main idea is that _you_ are always responsible for criticizing the links of proxies in your measurements. You must have an explanation for why the proxies relate to the quantity you are actually interested in. So _you_ are always grading your own homework! And thus must be honest and strive to seek the best explanations possible. --- Date: 20250104 Links to: Tags: References: * []()