## Question

Status:Closed Asked:Aug 16, 2015 - 12:32 PM

# Simple question about ATUS minutes to hours conversion

Hi everyone,

The time unit on the time use data in the ATUS is:

"DURATION is a 4-digit numeric variable reporting the length of the activity in minutes in one 24-hour period."

To convert it into hours per year, I would have thought to just multiply it by .164 since that's the fraction that converts minutes/day into hours/year. However, that generates values that are far from accurate in terms of time use within a year.

What simple observation am I overlooking?

To convert from minutes per day to hours per year, you would need to divide DURATION by 0.164, not multiply.

Hope this helps.

Based on your data extract, you are using the "ACT_" time-use codes. If I sum all of these codes together (a much broader definition than Aguiar and Hurst's "Leisure Measure 1"), the maximum value generated is 1,440 minutes, which corresponds to the total number of minutes in a 24-hour period. If you then apply the minutes/day to hours/year conversion factor of 0.164, you get a maximum of 8780.488 hours, which is slightly higher than the actual number of hours in a year. However, if you expand your conversion factor to more decimal places (or, just use the fraction 60/365), then you will get a more precise, and lower, maximum value. Assuming you have not double-counted any time-use categories, using a more precise conversion factor eliminates any cases of respondents reporting more leisure hours than the total number of hours in a year.

Hope this helps.

Based on your data extract, you are using the "ACT_" time-use codes. If I sum all of these codes together (a much broader definition than Aguiar and Hurst's "Leisure Measure 1"), the maximum value generated is 1,440 minutes, which corresponds to the total number of minutes in a 24-hour period. If you then apply the minutes/day to hours/year conversion factor of 0.164, you get a maximum of 8780.488 hours, which is slightly higher than the actual number of hours in a year. However, if you expand your conversion factor to more decimal places (or, just use the fraction 60/365), then you will get a more precise, and lower, maximum value. Assuming you have not double-counted any time-use categories, using a more precise conversion factor eliminates any cases of respondents reporting more leisure hours than the total number of hours in a year.

Hope this helps.

Thanks a bunch Tim for correcting my error. At about the 95th percentile, there are people using more time than there is in a year for Aguiar and Hurst's (2005) definition 1 for leisure (active and passive, socializing, gardening, general leisure). I can just drop them, but is this well-known?

Thanks again Tim for taking the time to reply and clarify! Some of the reported values are a little odd, and they get amplified when converting to hours/year. (For example, there are instances of 1440 minutes allocated completely to travel, work, and personal care.) Nonetheless, I just winsorize at the 99th percentile.

Appreciate the accessibility of IPUMS and your attention to detail on the questions!

### OTHER QUESTIONS NEEDING ANSWERS