New Zealand census data show a well-documented decline in mainline Protestantaffiliation and corresponding rise in non-religious identification from the mid-1960s onward.Despite extensive historical and sociological commentary on this transition, no formalstatistical test of the structural break has been published. We apply the Chow test exhaustivelyacross all candidate break years in the census time series for mainline Protestant share(Anglican + Presbyterian + Methodist), Catholic share, Pentecostal share, and No Religionshare. The optimal break for mainline Protestant decline is identified at 1971 (F = 97.4, p <0.0001), with the closely adjacent 1966 census also highly significant (F = 91.4). Before thebreak, mainline share declined at approximately 0.30 percentage points per year; after it, at0.86 pp/year — a 2.9-fold acceleration. The No Religion break is identified at 1976 (F = 169.7,p < 0.0001), one census later, consistent with a transitional lag. Results are confirmed bypermutation test (10,000 iterations, p < 0.0001 for all series). These findings provide the firstformal quantitative confirmation of what NZ religious historians have long describednarratively.