Quantitative methods make use of mathematical or statistical techniques in order to discern patterns about populations of interest. There are a variety of methods for undertaking quantitative monitoring and evaluation or impact evaluations. Household and individual-level data are typically collected using quantitative household surveys with a standardized questionnaire, typically with fixed coded responses, although some may allow open-ended responses to be coded later. Data for quantitative analyses may include panel data, that is data collected about the same households over a number of years, which allow for analysis of changes over time. Some of the surveys collect data at the level of the individual household member, which allows for comparison between men and women, and also helps to capture the full range of livelihood strategies within the household. Sampling to cover the range of wealth and/or poverty categories is critical for these types of surveys. Although some qualitative data is included in the quantitative surveys, researchers analyze most survey data—including qualitative responses—using statistical or econometric techniques in statistics packages such as SPSS, Stata, or SAS. For more detailed information about quantitative methods for impact evaluation and for monitoring and evaluation see the Methods section of the Gender and Assets Toolkit. For examples of quantitative survey modules that collect gender and assets data see the Quantitative Resources section.
Benefits
- Representativeness: Large sample sizes ensure that data will be more representative of the populations in question.
- Causation: Econometric methods allow you to test scenarios and calculate attribute causality, and estimate impacts to better understand which aspects of programs are more effective
- Availability of existing data. Some data is already publicly available in censuses and other databases thus it may be possible to conduct analysis without new data collection efforts.
Challenges
- Establishing context: With quantitative data it is more difficult to understand nuances of a given culture and context. This may lead to a tendency towards generalizations.
- Difficulties in establishing causality: Even with good data it is often difficult to establish causality.