The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » General » Agricultural Land ownership: Treatment of Zero values
Agricultural Land ownership: Treatment of Zero values Tue, 19 May 2020 17:28
 Indu Messages: 1Registered: May 2020 Member
Hello!

I am working on a project to understand the distribution of agricultural land ownership in different countries and have come across a lot of zero values for a few countries in the DHS, even while the HHSD reports yes to owning some land.

(example dataset: Bangaladesh: BDHR72FL.DTA)
The relevant variables of interest are hv244 (Binary Y/N response to owning agricultural land) and hv245 (variable capturing the area of land(ha.) owned if hv244 == Y). A basic (unweighted) tabulation of these variables gives us the following:

Year----Total sample HHSD------"Yes" to owning land----------Land (ha) == 0
2007----10,400------------------------4,823----------------- ---3,055
2011----17,141------------------------7,839----------------- ---4,857
2014----17,300------------------------7,735----------------- ---5,019

The table above shows that about 50% of the households who report yes to owning some land in hv244, report the area of land owned to be zero.

[Although the table here is only for Bangladesh, we have encountered a similar problem in other Asian (eg. Nepal) and Latin American (eg. Haiti) countries as well.]

What should be the right approach to correct for the situation and capture the area owned by such households?

I understand that the variable label for 2011 and 2014 (Bangladesh) specifies that values are given in one decimal so we divide other values by 10, but I am not sure how to treat these zero values.