[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
A.The assignment of anthropmetric z-scores based on the NCHS/CDC/WHO
International Reference Standard is done through a complicated interpolation
function that takes into account sex, age (measured by difference in date of
birth and date of interview, both precise to day of month), height in
centimeters, and weight in kilograms (precise to 100 grams). As part of a
creation of a recode file variables with the z-scores are calculated and
included in that file. In the process of assigning the z-scores, checks are made
on their plausibility. Z-scores are assigned missing to children with incomplete
date of birth (month or year missing or â??donâ??t knowâ??). The reason is because
z-scores are very sensitive to changes in age. Children with height for age
zâ??scores below â??6 SD or above +6 SD, with weight for age z-scores below â??6 SD or
above +6 SD, or with weight for height z-scores below â??4 SD or above +6 SD are
flagged as having invalid data. Also invalid are combinations of z-scores where
height for age is less than â??3.09 SD and weight for age is more than +3.09 SD,
or where height for age is more than +3.09 SD and weight for age is less than
B.The percentages of children stunted, wasted, and underweight are equal to the
specific numerators divided by the denominators and multiplied by 100.
Handling of Missing Values
Children who were not weighed and measured and children whose values for weight
and height were not recorded are excluded from both the denominator and the
numerators. Children whose day of month of birth is missing or unknown are
assigned day 15. Children who are flagged for out-of-range z-scores or invalid
z-scores are excluded from both the denominator and the numerators. Children
with missing data in the z-scores (date of birth incomplete or missing) are
excluded from both the denominator and the numerators.
If any one can help me out would be wonderful!!
Cheers / Tharshini
On 2009-06-14, at 11:45, Evans Jadotte wrote:
> Tharshini Thangavelu wrote:
>> Thanks for the help! It seems to be working to merge the data files.
>> For those who are familiar working with anthropometric data analysis, please
>> take a look at the question and clarify my argument. My question is about the
>> anthropometric data from the DHS. In the file the variables are defined as :
>> (HC70)Ht/ A Standard (according to WHO), (HC71) Wt/A Standard deviation (accod.
>> who). The values does not seem plausible. Here below you may find the summary
>> for the two variable. My question is, should I apply (manipulate the data and
>> recalculate with help of stata)according to the WHO growth standard? Since, it
>> can not be possible that a ghanian child is 145. 48 standard deviation behind a
>> well nourished child (the reference group)?
>> If the case is that I have to calculate my self according to the WHO growth
>> standard, how do I do this in Stata?
>> sum hc70 hc71
>> Variable | Obs Mean Std. Dev. Min Max
>> hc70 | 3419 -145.48 155.7705 -594 588
>> hc71 | 3419 -100.0962 118.903 -547 457
>> On 2009-06-12, at 14:00, Friedrich Huebler wrote:
>>> Tharshini has two files, one with anthropometric data for children and
>>> one with a listing of all household members that includes the same
>>> children. The data must be combined with -merge-, not -append-.
>>> On Fri, Jun 12, 2009 at 3:42 AM, Evans Jadotte<email@example.com> wrote:
>>>> Hi Tharshini,
>>>> Maybe you want to 'append' one file to another if what you need is to add
>>>> new variables to your master file.
>>> * For searches and help try:
>>> * http://www.stata.com/help.cgi?search
>>> * http://www.stata.com/support/statalist/faq
>>> * http://www.ats.ucla.edu/stat/stata/
> Indeed these numbers do not make sense (seen in standard deviations).
> Normally standard deviations in height z-scores and weight z-scores in
> the range
> -6 > height for age z-scores > 6
> -4 > weight for age z-scores > 4
> should be flagged as invalid. Check out this DHS' Guide to Statistics.
> Maybe you will get a clue for Ghana.
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
Forskarbacken 8 / 101
114 16 Stockholm
Phone +46 (0)735 53 43 90
* For searches and help try: