Different gridding methods provide different interpretations of your data because each
method calculates grid node values using a different algorithm. If you are not satisfied
with the map of your data, you might consider producing grid files using several different
gridding methods and comparing the results.
There are a number of different gridding algorithms, for example Inverse Distance to a Power, Kriging, Minimum Curvature, Modified Shepard's Method, Natural Neighbor, Nearest Neighbor, Polynomial Regression, Radial Basis Function, Triangulation with Linear Interpolation, and so on.
I have found the range of methods available in Surfer to be very impressive, allowing far more flexibility than many other (more expensive) gridding & contouring packages.