I think there are two parts to this. First, you select a model and have to make assumptions about your data and context to select an appropriate model. I tend not to think of these as assumptions of the model but statistical assumptions that always apply (e.g., that I haven't omitted important variables, that the structure of my data is captured by the model etc.).
Second, there are properties of the model itself, once selected. For linear regression these are typically assuming that residuals are sampled from an independent, normal population of errors with equal variance.
Other linear models might well have different assumptions (e.g., logistic regression).
There are some assumptions you need to follow, this link will help you out: https://www.youtube.com/results?search_query=todd+grande+linear+regression+assumptions
- Badges
- Science topic
- Similar topics
- Psychology