A focus on the examples that survive some process while accidentally overlooking those that did not survive — because they are no longer visible.
The example of using statistics of minimizing bomber losses to enemy fire in World War II was developed by statistician Abraham Wald, who noticed that they were only evaluating the the damage to planes that had survived, and thus skewed their analysis, as opposed to all bomber planes (and in particular, the ones that were shot down, did not return, and could not be studied for evaluation).
We watch entertainment that has survived many rounds of approval before it gets to us (we don't see all that fails). We think devices used to be built of higher quality, but only because the devices we encounter have lasted (and we forget those that have been discarded.) This type of bias creeps into our daily experiences from so many areas, it's useful to take stock when analyzing a group of something (items, processes, connections) and ask yourself about the examples that you might not be considering.