The Earth’s F region ionosphere is dominated by the collision between atomic oxygen and its first positive ion. An accurate corresponding collision frequency model is necessary to understand the ionosphere. However, the widely used classic Banks theoretical model typically provides a collision frequency that is 30% lower than the expectation from ionospheric observations. Accordingly, the classic collision frequency is often adjusted by multiplying it by a constant known as the Burnside factor. This correction-factor model adopted the classic model as its basis due to a misunderstanding that the classic model was based on a laboratory experiment; that is, the correction factor was originally meant to compensate for laboratory contamination. In this study, we construct a collision frequency model based on the laboratory experiment. We find that the resultant laboratory-based model is consistent with ionospheric observations. In this construction, we have determined that the impact of laboratory contamination is small (7%) and is mostly canceled by a misinterpretation regarding the conventional definitions of energy. Thus, the 30% difference is mainly caused by a theoretical error in the classic model itself. This error is energy-dependent and corrected by the later wide-energy theoretical model. Thus, the classic model cannot be corrected by a constant and should be replaced by the later model.