ABSTRACT Households in the immediate vicinity of sand, stone and gravel river mining all receive negative externalities, but some of them also benefit from mining activities. Differences in the perceived effects between community groups affect their acceptance of negative externalities. This study decomposes the calculation of the mean willingness to accept (MWTA) negative externalities from river mining activities between two community groups in the Batang Anai River area in the Padang Pariaman Regency, Sumatra, Indonesia: (a) farmers and (b) individuals involved in mining activities. First, the logit model is used to determine the effect of household characteristics on the probability of the community declaring environmental degradation. Next, the dichotomous choice contingent valuation method (DC-CVM) is employed to estimate the residents’ willingness to accept (WTA) the minimum value of compensation due to negative externalities from mining activities. We utilise four scenarios to prove the presence of bias if the decomposition is not taken into account in estimating the MWTA. According to estimation results, individuals who are engaged in mining have a 21% lower MWTA value than farmers. Meanwhile, indirectly affected communities who engage in mining have a MWTA value that is 32% lower than that of farmers. The implication of this study is that the decomposition of the MWTA by community groups is important to provide the real value of negative externalities.