The rate of mistakes by referees is way lower than the rate of mistakes by players. Just count the turnovers, fouls and missed shots per game. These go further to determining outcomes than any single foul call.
I had to laugh at the storm in a teacup over goal umpires' incorrect decisions in AFL this year. The stats were that goal umpires made 10 wrong calls in the whole season. That's out of about 10,000 decisions for the season - an error rate of around 0.1% . By contrast, the players' error rate (missed shots at goal) was around 50%. Yet people agonised about the desperate need to improve the quality of goal umpires' decisions. The cost per changed (improved) decision, would have been enormous. I see this discussion as similarly aiming to introduce a solution to a problem that will carry greater costs per changed decision than probably is warranted. The cost per changed game results (win/loss) would be even higher. This has to be a factor as we are dealing with a professional league that operates on business principles.
Likewise, in a professional league, more hangs on mistakes than in regular sport. Hence, players who make too many mistakes don't get played. Also, most people will continue to criticise refereeing as a single mistake by a referee is seen as more important than a single mistake by a player, at a given point in a game. Finally, we have the opportunity to have "do-overs" for referees' decisions but not so for players'.
That said, it is not sensible to assume the wrong call when the teams are neck and neck in the last 2 minutes will determine the game outcome. Maybe it was the wrong call on the previous play that resulted in a 4 point play, that put the teams neck and neck to begin with. Or a wrong call before that. Or the wrong call that put a key player off the court with their 5th foul and turned the game on its ear, 2 minutes into the third quarter. Or the wrong call that gave that player their first foul in the first minute of the game and meant they fouled out early in the last quarter. Or.....
So if we want to use video technology to review refereeing decisions, we need to recognise that limiting it to parts of games will just create pressure to use it for the rest of the games. So, I think it's all or none.
Nonetheless, I'm not sure I'd want to pay more to watch games just to have 1 or 2 refereeing decisions changed in each game.