This is an interesting poll. What they did was for half of their respondents, right before they were asked who they would vote for, they were asked "Would you say that your spouse earns more than you, less than you, about the same, or is your spouse unemployed?" For the other half, they were asked afterwards.
Turns out, this affected the Trump-Clinton results among men and women. For men, it changed the results from Clinton 49 - Trump 33 to Clinton 42 -Trump 50,or +16 for Clinton to +8 for Trump, or a 24 point swing to Trump. For women, it changed from Clinton 57-Trump 36 tp Clinton 59-Trump 26, or +21 to +33, or a 12 point swing to Clinton.
Keep in mind, the margin of error for thse subgroups is a little over 7%, so these numbers might be off somewhat, but there is clearly an effect among men.
The experiment was also done w/ regard to Sanders v Trump, but the effect there was less pronounced, with a swing of +1 among men for Sanders, and a swing of +4 among women.
(I'd have liked to see this done with a Sanders v Clinton poll)