And why should we trust some hamsters in a cage over the actual randomised control trials that have shown no statistically significant effect?
Not statistically significant is not the same as none.
If r have a rate of increase of 1.2 but masks reduce that rate by 0.5% and we start at 20 people Then after week one that's almost no difference, however after 15 weeks that's 308 who have been infected vs 286.
Even taking a lower figure and assuming a whole year of mask wearing at a 0.1% difference it's infections of 262,093 Vs 248,807.
If 0.1% die of those infected that's 13 deaths saved.
Hence why it has to be none, and not just not statistically significant. Few would argue that 0.1% is statistically significant yet for the friends and family of those 13 people it's enough of a difference.
That's because it's not just that you have to be protected once, but each and every time you interact with others. That also means that each time you stop one person from getting infected that's then a whole chain of people who aren't infected and therefore can't infect anyone else.
In one study quoted in this article:
The figures were:
1.3698630136986% N95
3.1185031185031% no protection
1.7486401048045% difference
3.2758620689655% surgical masks
3.9301310043668% no protection
0.6542689354012% difference
2.6422764227642% surgical masks
3.1185031185031% no protection
0.4762266957388% difference
I've used the percentages, as the size of the groups varied to the extent that in one example although more people got infected with masks the sample size was larger so overall there was still a decrease.
I would however add the notes that the group sizes were fairly small so there's likely to be a range (bigger/smaller difference) in the difference calculated as shown on the graphic, likewise it's not on Covid-19 which could result different figures and was within healthcare workers, if that was only whilst at work those being infected could have come from outside of their work settings (i.e. from their child from school), which again may not be the case for Covid-19.
As such the evidence isn't clear either way. Therefore because there might be some limited benefit it's worth trying.
However in saying that it needs to be highlighted that any such benefit is likely to be small and so masks shouldn't be seen as a replacement for other measures (good hygiene, staying 2m apart where possible, not mixing households inside unless well ventilated and for limited period, etc.). Unfortunately too many think that they are and so increase their risk more than would have been the case if they weren't wearing a mask.