Perhaps not exactly fifth-grade... but the reasoning is to maintain one significant digit in each number. Thus, 4.28 would be rounded to 4, and 4.28 million to 4 million.
When adding or subtracting, at what place you round (tens, hundreds...) is more relevant, but when multiplying or dividing, the number of significant digits is the relevant part.
>This is a sample problem out of a math workbook. I get that "estimating" essentially means performing calculations on rounded values, but they are mixing the places they round to in the example below. Why isn't it either 1 x 4 or 0.8 x 4.2? Not quite getting how they are doing that. I thought with rounding, you had to know the place you wanted to round to in order to have the product be valid.
>
>
>Estimate Decimal Products
>-------------------------
>
>Estimate 4.2 x 0.843
>
> 0.843 --------> 0.8
>x 4.2 --------> x 4
> ====
> about 3.2
>
>Both factors are rounded
>down. The actual product
>is greater than 3.2.
>
Difference in opinions hath cost many millions of lives: for instance, whether flesh be bread, or bread be flesh; whether whistling be a vice or a virtue; whether it be better to kiss a post, or throw it into the fire... (from Gulliver's Travels)
