This paper considers the case for replacing the Carli index in the Retail Prices Index for calculating price changes at the elementary aggregate level. Following Diewert (2012), we go through each of the three approaches used to select appropriate index numbers: the test, stochastic and economic approaches. In each case, we find a few areas where our conclusions differ from Diewert's. Unlike Diewert, we are not as concerned that the Carli fails the time reversibility test, but note that it fails a revised price bouncing test. We find that the stochastic approach is inapplicable at the level of elementary aggregates, where by definition quantity weights for goods are unknown. However, we argue using insights from information theory, that the economic approach can be applied at this level and moreover that it favours the use of the Jevons index.