10 vs 1 is not that insignificant. If you expect deviation of sqrt(n), this is about 2 sigma, i.e. 95% significance..
WN and UA will each do over a million landings a year. Are you saying that 1 in a million and 10 in a million is a statistically significant difference? That would surprise me.
Sure it is. It is a whole order of magnitude. 1 in a million or 1 in 100,000. 1 overrun every decade or 1 a year. Big difference.
What @sgbroimp is saying, is that for small number of events it can be more about luck than anything else.
Is airline with 0 crashes over 10 years really safer that airline with 1 crash over same period? Nope, chances are they are just more lucky.
What about 1 crash and 2 crashes? Same thing. 1 and 5? well, probably there is a trend, but luck still not out of question. Official review is warranted, though. 1 and 10? still small chance of bad luck (5% maybe) - but 95% chance that something is wrong and someone has to take an action. 1 and 100? Miracles happen, but nobody believes in them.
That is why serious accidents, major crushes, hull losses - which are becoming rarer by the year - are not a good measure of safety practices. Small scale events, inspection results, close calls - those are a much better metric, there are more of those, and they paint a better pictire. Although that maybe not public data, and media is not willing to go into that. Safety-oriented management, as well as oversight agency, must keep an eye on that data. And so do we, if we want to keep a.net as a quality discussion space.
That is why I saying "I don't fly Southwest because of it!" is just not so smart. Either dig up more data and build a better case - or feel free to pay more for less, industry needs free cash. And I, for example, would prefer WN over SQ any time, as SQ has a history of interesting events being treated in an interesting fashion... but that is getting off topic.