Question

Status:Closed    Asked:Jun 06, 2017 - 05:15 PM

Codes for the BMI variable

Below are the codes for the Body Mass Index (BMI) variable (not the bmicalc) on IPUMS .

0000: NIU

9980: 99.8+
9999: Unknown


I thought that those with the BMI of 99.8 and over were coded as 9980.

However, when I tabulate the BMI variable (across several years before and 10 years after 1997), I found that 83 people for 99.90 and 2,586 people for 99.99 and there were no observations for 9980 and 9999. Are there different codes for the variable?

Please see the tabulated results below.


Thanks for your answer in advance!



Variable | Obs Mean Std. Dev. Min Max

-------------+---------------------------------------------------------

bmi | 156,878 13.61521 17.91787 0 99.99

. tab bmi if bmi ==99.99

no observations

. tab bmi if bmi ==99.90

no observations


. tab bmi if bmi >99.8

Body mass |

index | Freq. Percent Cum.

------------+-----------------------------------

99.90 | 83 3.11 3.11

99.99 | 2,586 96.89 100.00

------------+-----------------------------------

Total | 2,669 100.00




 
Do you have the same question? Follow this Question
 

Staff Answer

avatar

Joe_Grover

Staff

The BMI variable is stored as a four digit number with two implied decimals (as stated in the BMI codes tab). This means that when you download the data in a fixed-width format the BMI values range from 0001 to 9999. When the data is then read into a stats package, such as Stata or SPSS, the syntax file provided by IPUMS includes code to divide in the implied decimals so the range of values for BMI becomes 00.01 to 99.99

The 99.90 values you are seeing are all from the 1992 sample and should be treated as 99.99 (NIU). The IPUMS Health Surveys team is aware of the issue and it will be corrected with the next data release.

I hope this helps. I will let the IPUMS Health Surveys team know that this was not clear and hopefully we can develop a more clear description.

 

Jun 07, 2017 - 11:32 AM

0
0
Report it

OTHER QUESTIONS NEEDING ANSWERS

negative values of wtsupp 1975 and before
How to link original ASEC family relationship variable with RELATE variable o...
For the variable BMI, why are 65% of cases not in universe for the combined 2...
Why are not all the "not in the labour force" categories filled for the EMPST...
"Years spent in the US (YRSINUS)" variable in 1998
Login   |   Register


Share |