This paper presents the application of non-probability sampling – the rivers sample method to reach a group of professional foster parents in Poland – case of hard-to-reach population. This is a population whose public, nationwide register does not exist and additionally, it is kittle known about it in terms of the characteristics of its members, therefore quota sampling was not feasible. River sample method involves recruiting respondents by inviting them to the survey while they are doing some other activity on the Internet. The invitation to take part in the survey can be made through online banners, advertisements, information in forums or e-mails and on thematic websites. It provides a quick, low-cost, flexible way of reaching the group of respondents who are the target but it has its limitations- under-coverage and self-selection – which may affect the bias in the sample and thus the possibility to generalise to populations. Despite those limitations the presented study obtained a sample that reflects the proportions regarding the known characteristics in the population of professional foster carers in Poland: their provincial distribution and diversity of forms. In general, the river sample method works efficiently especially for exploratory studies and hard-to-reach groups.