In his book Differential Forms with Applications to the Physical Sciences, on pages 19–20, Harley Flanders writes:
"a one-form $$ \omega = P\,dx+Q\,dy+R\,dz$$ may be identified with an ordinary vector field $(P,Q,R)$ in $\mathbf E^3$, a two-form $$ \alpha = A\,dy\,dz+B\,dz\,dx+C\,dx\,dy $$ may be identified with a polar vector field in $\mathbf E^3$."
What do "ordinary" and "polar" mean here?
Physicists tend to define geometrical entities by how they change under a coordinate transformation. Those transformations are almost always rotations and parity, that is, reversing all directions (in 3 dimensions, this reverses the orientation too). If something "rotates like a vector", physicists call it a vector, no matter what.
Now in a Euclidean setting, if you only look at rotations, you can't really distinguish a vector from a 1-form, because they change with the same transformation. If a vector transforms with the linear map $A$, a 1-form transforms with the map $(A^t)^{-1}$, which for rotations is equal to $A$. The very same can be said about 2-forms in 3 dimensions.
Under parity, 1-forms and vectors transform the same way: they simply reverse. 2-forms, instead, are invariant, they don't change sign.
Now, before the discovery of differential forms (and bivectors), physicists already had to use something that rotates like a vector but is invariant under parity. This was needed, for example, for angular momentum, or for any vector that was the vector product of two usual vectors (can you see why?). Such a "vector" is called pseudo-vector, or axial vector, or polar vector.
The magnetic field is the nicest example of a "polar vector" that can be modelled better by a 2-form.