The goal of the text is to analyze concordance between one’sreligion and a close relative’s estimation thereof. To establish the accuracy of parents’ estimates of their children’sreligion and vice versa, we ask the following question: what is the concordance of post-socialization beliefs about (un)successful transmission of religiosity between direct actors (parent/child)? We argue that the reliability of that estimate indicates the effectiveness of religious socialization. Socialization is not treated as a nonproblematic one-way process, but rather as a result of repeated mutual parent-child interactions and a host of other intervening factors (secondary socialization etc.) In this context, the level of estimate reliability is treated as an indicator of religious socialization and of the continuity of religious memory within family, which is viewed as a collective phenomenon. In other words, by imprinting values into one’s memory and worldview, the process of religious socialization shapes the ways one views the world and him/herself as well as the focus of his/her attention, or what is stored in his/her memory. Our project is conceptualized at an intersection of the theories of socialization and religious memory. Among the latter, we primarily rely on Jan Assmann’s conceptualization of memory. While many contemporary authors deal with issues of religious socialization, and some even with its links to memory, no investigations thus far have attempted to verify intergenerational transmission in terms of the reliability of mutual estimation of (non)religiousness between generations. The analyses are based on data from the KODINA 2015 survey which specifically complemented its main sample with a selection of respondents’ close relatives. This allows us to analyze a sample of parent-child dyads. Simple frequency tables revealed some effects of the accounting person’s generation on estimate concordance. Specifically, parents were much more likely to misjudge their children’s religiosity - to consider them Roman Catholic although the children themselves identified as nonbelievers. Logistic regression was used for more detailed analysis, which revealed that the discordance between parents’ and children’s estimates is not statistically significant and only a partial effect could be demonstrated in interaction with the estimating person’sreligiosity. In sum, we present three main findings: (1) close relatives estimate each other’sreligiosity relatively accurately. That indicates, as we argue, the importance of the socialization process. (2) Estimate accuracy depends on the estimating and estimated persons’ religiosities and combination thereof. This demonstrates the importance of continuity of (non)religious belief because lower estimate accuracy has been observed in discontinuity situations. (3) There was no statistically significant difference of estimates between generations, yet the relationship demonstrated in the case of the interaction is primarily notable in that Roman Catholic parents incorrectly estimated their nonbeliever children to be Roman Catholics as well.