Difference between format and value

Can we clearly make the difference between data format and data value in any kind of language (in a theoretic point of view?) or is the value of a data a sub-set of its format?

My objective here is to identify if an incoming data for a given function is correct. An instinctive way to do this is just to check the value you expect. But if the data is a structure (or a complex object), you will also check (in the way you want) if the structure or object received is the one you expect (with the correct format and the correct values).

An other good example is those of a signal where the shape (format) is different, but the values are in the same boundaries.

So do you guys think that these separation between data format and data value is something that make sense in a verification (for a critical system for example) point of view? Can we always make this separation or is it sometimes self included?