The three main methods currently in use for estimating the excitation class of planetary nebulae (PNe) central stars are compared and evaluated using 586 newly discovered and previously known PNe in the Large Magellanic Cloud (LMC). In order to achieve this we ran a series of evaluation tests using line ratios derived from de-reddened, flux-calibrated spectra. Pronounced differences between the methods are exposed after comparing the distribution of objects to their derived excitation. Line ratio comparisons show that each method's input parameters have a strong effect on the estimated excitation of a central star. Diagrams were created by comparing excitation classes with Hβ line fluxes. The best methods are then compared to published temperatures using the Zanstra method and assessed for their ability to reflect central star effective temperatures and evolution. As a result we call for a clarification of the term ‘excitation class’ according to the different input parameters used. The first method, which we refer to as Exneb relies purely on the ratios of certain key emission lines. The second method, which we refer to as Ex* includes modeling to create a continuous variable and, for optically thick PNe in the Magellanic Clouds, is designed to relate more closely to intrinsic stellar parameters. The third method, we refer to as Ex[Oiii]/Hβ since the [Oiii]/Hβ ratio is used in isolation to other temperature diagnostics. Each of these methods is shown to have serious drawbacks when used as an indicator for central star temperature. Finally, we suggest a new method (Exρ) for estimating excitation class incorporating both the [Oiii]/Hβ and the HeII λ4686/Hβ ratios. Although any attempt to provide accurate central star temperatures using the excitation class derived from nebula lines will always be limited, we show that this new method provides a substantial improvement over previous methods with better agreement to temperatures derived through the Zanstra method.