The Si/Al contact resistance of narrow p-type contact areas of crystalline silicon solar cells is a crucial parameter for the further development and improvement of passivated emitter and rear (PERC) solar cells. In this paper, two methods for determining the rear contact resistance are analyzed and improved: the series resistance analysis and the transfer length method (TLM). We use an experimental set of PERC solar cells with varying metallization fraction, and we analyze the experiment using numerical device simulation. In the first method, the resistive losses are extracted from the I-V-curves of the PERC cells, and the rear contact resistance is separated by subtracting the analytically calculated base contribution. The state-of-the-art analytical calculations deviate significantly from our numerical simulations, causing a high level of uncertainty that we do not recommend this method (an upper limit of the rear contact resistance per line contact is determined to be about 2 Ω cm). In the second method, the rear line contacts are separated and measured with the TLM. Still, the series resistance loss in the base has to be quantified considering a thick base and the rear line contact geometry. We separate the base component by numerical device simulations and find that the rear contact resistance per line contact is below 0.5 Ω cm. This implies that rear contact resistance reduces cell efficiency of screen-printed PERC solar cells only by about 0.1%abs for metallization fractions higher than 3.5%, i.e., the rear contact resistance is currently no significant loss mechanism in our PERC solar cells. All of our improved evaluations lead to lower resistance values than previously reported and with a reduced uncertainty. The TLM method is suggested to be favorable.