#science One of the things you'll learn if you go through a scientific PhD is the value of models. Models are representations of real-life situations that we simplify in order to gain insights into those situations. Some examples include mathematical models of how infectious diseases like Covid spread, and ideological models of human personalities like the big five personality traits. These models have a representation of how a system or person behaves, and we use that representation to predict the outcome of a situation. The foundational thing I learned in grad school is that _all models are wrong, and some models are useful_. It's crucial for us to understand what a model does and does not accurately represent, because from that we can identify the things that the model will accurately predict, and the things it will inaccurately predict. Being able to quickly understand what information is accurate and what information is not boils down to understanding 1) the model being used to predict that information and 2) the assumptions about the data being input into the model. The second point is much easier to understand, so let's just get that out of the way quick. If the data being input is not consistent with the assumptions of the model, the model will not predict an accurate answer. As a quick example of how we do this every day, take a model of dog training that might say: "when an owner trains their dog to sit by repeating the word _sit_ over and over while teaching the dog to sit, the dog associates the word _sit_ with sitting and typically follows that command in the future". This is a great model of dog training; but if someone tries to apply this model to a deaf dog, the prediction that the dog will follow the _sit_ command in the future will fail miserably. That example is obvious, and I chose such an obvious example to demonstrate that we perform this "model + data" analysis in our heads every day to nearly every situation. The better we get at analyzing these models, the better we get at predicting outcomes - and the best way to get better at analyzing these models is to understand their fundamental parts. The fundamental parts of every model are 1) the assumptions about the input information and 2) how the model does its prediction. Couple that with the understanding that _all models are wrong, and some models are useful_ and you can quickly build the skills to have an incredibly nuanced understanding of how the world works, from the spread of diseases to human behavior.