Difference between revisions of "Secondary Data Sources"

Jump to: navigation, search
(5 intermediate revisions by one other user not shown)
Line 1: Line 1:
<onlyinclude>
<onlyinclude>
Impact Evaluations rely on many different sources of data: administrative, survey, geospatial, sensors, telecomms, and crowd-sourcing. An important step in designing an impact evaluation is to evaluate what data sources are best suited (and which are available, given the context).
Secondary data is data collected by any party other than the researcher, including administrative data from programs, geodata from specialized sources, and census or other population data from governments. Secondary data provides important context for any investigation, and in some cases (such as administrative program data) it is the only source which covers the full population needed to conduct a research project.
</onlyinclude>
</onlyinclude>
Impact Evaluations rely on many different sources of secondary data, such as: administrative, geospatial, sensors, telecomms, and crowd-sourcing. An important step in designing an impact evaluation is to evaluate what data sources are best suited (and which are available, given the context).


== Guidelines ==
== Guidelines ==
Line 9: Line 11:
[https://www.povertyactionlab.org/ JPAL] provides a [https://www.povertyactionlab.org/na/administrative-data-and-evaluation-guides useful guide to using administrative data for impact evaluations].
[https://www.povertyactionlab.org/ JPAL] provides a [https://www.povertyactionlab.org/na/administrative-data-and-evaluation-guides useful guide to using administrative data for impact evaluations].


=== Survey Data ===
=== National Survey Data ===
The bread and butter of most impact evaluations is primary data collection; enumerators conducting personal interviews with respondents. These can be in the form of household surveys, firm surveys, school surveys, health facility surveys, etc. They can take place in-person, by telephone, or online.  
Existing survey data may be of use depending on the sampling frame for the impact evaluation, level of representativity of the existing data, and availability of disaggregated data. National Statistics Office typically collect a wide array of nationally-representative data, such as Living Standards Measurement Surveys and censuses. International survey efforts such as the Demographic and Health Surveys [https://dhsprogram.com/] and Enterprise Surveys [http://www.enterprisesurveys.org] are also good sources.  


=== [[Geo Spatial Data]] ===
=== [[Geo Spatial Data]] ===
Line 19: Line 21:


=== [[Telecom Data]] ===
=== [[Telecom Data]] ===
This includes [[call detail records]], social media data, web scraping.
This includes call detail records, social media data, web scraping.


=== [[Crowd-sourced Data]] ===
=== [[Crowd-sourced Data]] ===
Line 25: Line 27:


== Back to Parent ==
== Back to Parent ==
This article is part of the topic [[Data Sources]]
This article is part of the topic [[Secondary Data Sources]]




Line 31: Line 33:




[[Category: Data Sources ]]
[[Category: Secondary Data Sources ]]

Revision as of 14:50, 12 February 2018

Secondary data is data collected by any party other than the researcher, including administrative data from programs, geodata from specialized sources, and census or other population data from governments. Secondary data provides important context for any investigation, and in some cases (such as administrative program data) it is the only source which covers the full population needed to conduct a research project.


Impact Evaluations rely on many different sources of secondary data, such as: administrative, geospatial, sensors, telecomms, and crowd-sourcing. An important step in designing an impact evaluation is to evaluate what data sources are best suited (and which are available, given the context).

Guidelines

Administrative and Monitoring Data

Administrative data includes all data collected through existing government Ministries, programs and projects. It is a potentially rich source of data for an impact evaluation. Key challenges are: data is in paper format only (needs to be digitized), restricted access, lack of numeric identifier (or lack of common identifier with other key datasets).

JPAL provides a useful guide to using administrative data for impact evaluations.

National Survey Data

Existing survey data may be of use depending on the sampling frame for the impact evaluation, level of representativity of the existing data, and availability of disaggregated data. National Statistics Office typically collect a wide array of nationally-representative data, such as Living Standards Measurement Surveys and censuses. International survey efforts such as the Demographic and Health Surveys [1] and Enterprise Surveys [2] are also good sources.

Geo Spatial Data

This includes data from traditional satellites, micro- and nano-satellites, and unaccompanied aerial vehicles (UAVs, e.g. drones).

Remote Sensing

This includes all data collected by sensors, and through the Internet of Things (IoT).

Telecom Data

This includes call detail records, social media data, web scraping.

Crowd-sourced Data

This includes all data collected by crowd-sourcing, often through social media or mobile apps.

Back to Parent

This article is part of the topic Secondary Data Sources


Additional Resources