Don't judge a book by its cover. That's a wise man say. Never judge something by just looking at a glance. Data user from all over discipline must know about this. When they have data, they have to recognize where and how it was collected, not specifically, but they have to know.
Statistics Indonesia is an official institution that provide macro data such as number of poor household, The data itself speaks not only to show where the position of Indonesia in the world, but also used to draw some policies. It is strategic but also tricky in term how the data is used.
I have this information from a friend who responsible for the methodology on how the data was collected and smoothed. After they collected data from the field (they use the term field to represent where the respondents is located), it was weighted and judged by several variables, But not only that. They have to be smoothed because there are many political aspects on this. That's why there is one more process called proxy mean test.
Proxy Mean Test is counted with the formula Ln (Yi) = BX + Err(i) The model is supported with Susenas data from 2003-2008 and Podes (2 point of years). Although the reability of those two kind of data is still questioned by data user, but those data is the best we have (in Indonesia). There is no other institution can provide data like that (with lower budget)
Say, you have only couple of billion to provide micro data to draw policies that affect the whole country. The difficulties of collecting data is high but the pay off is low. What would you do? Science takes it all! And proxy mean test is one of the methodology that help Statistics Indonesia to provide data. Pretty smart eh..
>> I posted this about 2 years ago in my multiply
No comments:
Post a Comment