I guess you're referring to the comment on marriage or common-law relationships playing a role in people exiting low income. We basically use longitudinal data where we track individuals through time. We observe a single person who's in low income, and we look at the probability of their exiting once they're married. We can see the effect of the marriage or the common-law relationship on their income, moving them out of low income.
On February 26th, 2009. See this statement in context.