Scientific notation

[Mohsin Javed]

I have an individual level dataset with around 11 million observations. For each individual, I have data on individual's occupation (occ), hours worked in occupation (hours) and weeks worked (weeks).
I create a variable which measures labour supply as labour_supply = weeks * hours
Next, I find the total labour supply by occupation using the command: collapse (sum) labour_supply, by(occ)
However, in doing so, I lose precision as STATA shows the labour supply by occupation in scientific notation. Moreover, labour supply by occupation is shown with varying precision levels. For example, I get 1.41e+09 for occupation X and 5.02e+07 for occupation Y.
Is there a way around to get the exact values for this variable?

[Jared Greathouse]

Please show your dataex example

[William Lisowski]

If you have numbers on the order of 10⁹ then you have possibly lost precision in your calculations by not using double or long as the storage type for labour_supply. Then, assigning an appropriate format will handle displaying the results in their full precision.

generate double labour_supply = weeks * hours
format %12.0fc labour_supply
collapse (sum) labour_supply, by(occ)

[Jared Greathouse]

Another solution to this is just divide by a billion. Then your results will be the exact same, just as in billions.

So if your OLS coefficient is 5.8, a one unit increase in x means a 5.8 billion increase in the outcome variable when we've adjusted for other predictors. Then you don't gotta go through this problem.