Yes. In fact, people are so likely to make assumptions about future performance based on past performance that we invented a word for that social construct: "reputation."
Yes, because it's easier to do that than assess people on their merits. This is especially true when you are reviewing a technical paper (often for free) when you have competing demands on your time and mental energy (research, teaching). It takes hours and hours to properly review a paper, and that is still trusting that the author is acting in good faith - to reproduce their work might in some cases require a large amount of time and money. These are the same reasons that letters of recommendation (written by people you've heard of) are so useful in academia. Lots and lots of bias problems, but the bias isn't the point, as I think is sometimes implied.
