Consistent in two elections? That’s not consistent. That’s not even data, let alone a trend.

As I said, pollsters adjust the demographic weighting based on election results. It is possible they will again underestimate Trump’s performance. It’s also possible they will overestimate it. Only time will tell.

But regardless of that issue, it is within the margin of error—that is a statistical reality irrelevant to your speculation about polling errors.