Re: Basic statistics question

On Jul 14, 1:33?pm, Person <facetious_nickn...@xxxxxxxxxxx> wrote:
OK, hypothetically, if I'm keeping track of my expenses, is there a
statistically meaningful way of determining the effect that rounding
towards the nearest dollar will have on the accuracy of the
record-keeping? Upon what variables would this reckoning depend?

Expressed plainly, if I record $1.49 as $1.00 and $1.50 as $2.00,
will the roundings eventually cancel one another out?

Perhaps this question cannot be answered as asked?

Thanks, regardless.

If you round up and round down to the nearest dollar, the error in the
sum diminishes as the length of the column of figures grows. In other
words, if you create 2 columns of figures in a spread*** with 20
entries in each column using the actual figures in one column and the
rounded up or down figures in the other column, you will have a
certain percentage of error between the two sums. Now enter 40
figures, still rounding up or down only in the one column, and the
percentage of error between the real sum and the rounded sum will be
reduced. More data points lessens the error. However you will seldom
achieve zero error.

.