Technically, the margin of error of a poll, in this case +/- 4%, would mean the actual populace might be either 4 points higher or 4 points lower than the reported 7% lead. So it could be as low as a 3% lead or as high as an 11% lead. So that's an 8 point range within which the actual voters might fall.
Notably, a study reported in the New York Times in 2016 found that most political polls actually have ended up with about a 6 to 7% margin of error compared to actual voting results. (That's a range of 12 to 14 points total between the minus end and the plus end.) So you can't always assume that polling results will predict actual election results.
To multiply three factors, we first multiply two of the factors together. Then we multiply the product we get by the third factor. Since multiplication is commutative, we may multiply numbers in any order.
