In atomic force microscopy (AFM), typically the cantilever's long axis forms an angle with respect to the plane of the sample's surface. This has consequences for contact mode experiments because the tip end of the cantilever, which is constrained to move along the surface, displaces longitudinally when the applied load varies. As a result, the AFM tip makes contact with a different point on the surface at each load. These different positions lie along the projection of the lever's long axis onto the surface. When not constrained by static friction, the amount of tip-displacement is, to first order, proportional to the load and is shown to be substantial for typical AFM and cantilever geometries. The predictions are confirmed experimentally to within 15% or better. Thus, care should be taken when performing load-dependent contact mode experiments, such as friction versus load, elasticity versus load, or force versus displacement measurements, particularly for heterogeneous or topographically-varying samples. We present a simple method to reliably and precisely compensate for in-plane tip displacement that depends only on the range of vertical motion used to vary the load. This compensation method should be employed in any load-varying AFM experiment that requires the tip to scan the same line or to remain at the same point at each load. ©2005 American Institute of Physics