Take the learning further by asking the child which estimate is closer to the exact sum, rounding to the nearest ten or front-end estimation. Why does he/she think that is so? (front-end estimation will consistently fall behind). Is there a way to adjust estimation for 2-digit sums to make the process better?

Front-end estimation keeps falling behind the exact sum, while rounding is sometimes behind and sometimes ahead. Using this problem as an example: 71 + 38 + 56 + 10 + 43 + 37 + 22 = ?

--Add a 5 to the front-end estimate for every number in the sum, and use the Look for Tens strategy to add this to the front-end estimate. Rounding to the ones place makes a number estimate that is too large by up to 5 or too small by up to 4. Front-end estimating makes an estimate that is never too large, but can be too small by up to 9. If you add 5 to every front-end estimate, the estimate is now too small by up to 4 or too large by up to 5--the new estimate is sometimes ahead and sometimes behind, just as it is for rounding.