Background The similarity of friends in the frequency and quantity of alcohol consumption is explored. Method During their first semester, 57 psychology freshmen indicated weekly drinking frequency and quantity and nominated the three peers of this group they liked most. These nominations were then used to derive the weekly alcohol consumption of friends that either did or did not reciprocate a nomination. Results Multilevel modeling of weekly variations showed that individuals' drinking frequency was similar to peers who reciprocated a friendship (b = 0.15, p = .001), but not to non‐reciprocating peers (b = −0.01, p = .720). In contrast, weekly variation in quantity of individual students' drinking was similar to both reciprocating (b = 0.11, p = .018) and non‐reciprocating peers’ drinking (b = 0.10, p = .014). Yet across all weeks, quantity tended only to be similar to non‐reciprocating peers (b = 0.49, p = .020). Conclusions Freshmen might spend drinking time with peers who reciprocate a friendship, but are similar regarding the quantity of drinks consumed to all people they find interesting. Thus, alcohol consumption is used strategically for social purposes. This social purpose should also be acknowledged in alcohol‐reduction interventions.