Glucose testing consent [message #16589] |
Mon, 04 February 2019 10:11 |
DHS user
Messages: 111 Registered: February 2013
|
Senior Member |
|
|
I am currently using the data retrieved from the DHS website for my study. I am applying a multilevel model to the data . I am using individuals s my 1st level, households as 2nd level and am hoping to use districts as my third level. Is there any way I can access information that is captured for districts?
I am focusing my research to the state of Kerala. I have seen that consent for glucose testing (variable shb20) is missing for around 32,000 observations. This variable has been left as missing ( not refused/not granted). Variables pertaining to height, weight, Haemoglobin, SBP and DBP are also missing for these 32,000 observations.
Was consent asked for all individuals in the dataset? If no, is there a way to identify how these individuals were identified as eligible to ask for consent? Can I assume that data is missing, i.e. consent was asked but not recorded for these individuals?
Please advise on how to proceed. Hope to hear from you soon.
|
|
|
Re: Glucose testing consent [message #16590 is a reply to message #16589] |
Mon, 04 February 2019 10:12 |
Bridgette-DHS
Messages: 3199 Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Specialist, Fred Arnold:
In NFHS-5, a subsample of household was selected to be part of the state module The state module includes a random selection of every other household in 30% of the primary sampling units. The remaining households are in the district module. The district module includes the entire woman's questionnaire and all of the biomarkers (the CAB component of the survey). The state module excludes the biomarkers and Sections 8-11 of the woman's questionnaire. There is a state module variable in the dataset which indicates whether a household is in the state module. Since most households are not in the state module, most households were not eligible for the biomarker tests. That's why you're seeing so many missing values for the biomarker tests.
|
|
|