In this study, the accuracy of predicted psycho-socio-economic impacts resulting from mining and infrastructure projects is investigated by comparing it to impacts shown to actually occur. A systems theoretical approach was followed in conjunction with a qualitative methodology in order to conceptualise impacts in the social systems they occur. Data was collected through a document review (which included a total of 17 documents pertaining to predicted impacts, and 24 documents pertaining to actual impacts) and analysed by means of thematic analysis, which rendered four main themes and 20 subthemes. The findings of the thematic analysis were subjected to second-order analysis, which enabled the categorisation of impacts according to the level of accuracy with which they are predicted. To understand why some impacts are incorrectly predicted, a third-order analysis was performed. The study suggests that many of the commonly predicted psycho-socio-economic impacts are less accurate than what they should ideally be, suggesting that some of the assumptions on which these predictions are based should be revised, as should the conceptualisation of the impacts. The researcher argues that, in order to make accurate predictions about the impacts resulting from mining and infrastructure projects, sufficient knowledge of the attributes of the project, the nature of the receiving environment, the causal processes by which the project will bring about changes in the receiving environment, and the value systems according to which communities judge whether a specific change constitutes a negative or positive impact, is required. This argument is substantiated by highlighting instances of inaccurate predictions relevant to each category of required information. More fundamentally, however, the researcher argues that inaccurate predictions are the result of inadequate consideration of the systemic nature of psycho-socio-economic impacts and the context in which they occur, precipitated by the incorrect use of the “social impact” metaphor.