I'm basing this question on Fowler PoEAA. Given your familiarity with this text, aren't the ViewModels used in ASP.NET MVC the same as DTOs? Why or why not? Thank you.
They serve a similar purpose (encapsulating data for another layer of the application) but they do it differently and for different reasons.
The purpose of a DTO is to reduce the number of calls between tiers of an application, especially when those calls are expensive (e.g. distributed systems). DTOs are almost always trivially serializable, and almost never contain any behaviour.
For example, you're developing an e-Commerce site. CreateCustomer and AddCustomerAddress are separate operations at the database level, but you might for performance reasons want to aggregate their data into a NewCustomerWithAddressDto so that your client only needs to make one round-trip to the server, and doesn't need to care that the server might be doing a bunch of different things with the parcel of data.
The term "ViewModel" means slightly different things in different flavours of MV*, but its purpose is mainly separation of concerns. Your Model is frequently optimised for some purpose other than presentation, and it's the responsibility of the ViewModel to decouple your View from the Model's implementation details. Additionally, most MV* patterns advise making your Views as "dumb" as possible, and so the ViewModel sometimes takes responsibility for presentation logic.
For example, in the same e-Commerce application, your CustomerModel is the wrong "shape" for presentation on your "New Customer" View. For starters, your View has two form fields for your user to enter and confirm their password, and your CustomerModel doesn't contain a password field at all! Your NewCustomerViewModel will contain those fields and might, depending on your flavour of MV*, be responsible for some presentation logic (e.g. to show/hide parts of the view) and basic validation (e.g. ensuring that both password fields match).