As a North American social scientist, I would probably have reservations with accepting data from a South African study. I say "probably" because to some degree this depends on what the study is about. However, for any study that is specifically relating to North America, the South African data may not be reliable.

This is primarily due to socio-economic, political and historical differences between the US and South Africa. The challenges faced by South Africa are very different in most cases to those faced by Americans, differences which may manifest in the study.

People from many cities in the US would take things like public transport or the ability to go to a state hospital for granted, which leads to a very different way of life and different ways of thinking.

That said, it would depend on what the study was about. If, for example, you were doing a worldwide survey on whether people preferred Italian or Mexican food, then it would not matter where you respondents were from or what their background was. In most cases, however, I would argue that the respondents to any sociological survey should be from the geographical location and socio-economic group related to the study.