The presence of hopping carriers and grain boundaries can sometimes lead to anomalous carrier types and density overestimation in Hall-effect measurements. Previous Hall-effect studies on carbon nanotube films reported unreasonably large carrier densities without independent assessments of the carrier types and densities. Here, we have systematically investigated the validity of Hall-effect results for a series of metallic, semiconducting, and metal–semiconductor-mixed single-wall carbon nanotube films. With carrier densities controlled through applied gate voltages, we were able to observe the Hall effect both in the n- and p-type regions, detecting opposite signs in the Hall coefficient. By comparing the obtained carrier types and densities against values derived from simultaneous field-effect-transistor measurements, we found that, while the Hall carrier types were always correct, the Hall carrier densities were overestimated by up to four orders of magnitude. This significant overestimation indicates that thin films of one-dimensional SWCNTs are quite different from conventional hopping transport systems.