A matched pairs design is treated as one sample for intervals and testing purposes despite there being two samples taken because it is taken as a pair and not taken individually. Although the results might be different for each design, the matched pairs will represent one result.