I would like to know which attributes can be used to easily identify these two forms of data sets
The basic difference is dimensions. Panel data has three dimensions, other has two.
Combination of cross sectional data(data collected on several
individuals/units at one point in time) and time series data(data collected on one individual/unit over several time periods) is panel data. Panel data are repeated cross-sections over time, in essence there
will be space as well as time dimensions.
thanks a lot Ayodele for the opinion shared. What in your opinion then is a pooled cross-section?..Has it no space and time dimension?
David,
I'd go less technically:
As Ayodele said, panel data are collected on a cohort, or panel, of the same respondents repeatedly over time, and cross-sectional data are collected on each respondent only once.
"Pooled cross-section" would typically be on respondents of different "loci" of time (usually age, but might be time-from-relapse or any number of things) but without measuring the same respondents repeatedly.
In either case, age can be used as a predictor, but only with panel data can intra-individual change be assessed. There are also different implications for birth cohort vs. calendar time as predictors. It's also worth noting in some research domains that panel data are very difficult, though not necessarily impossible, to collect anonymously.
Pat
oh ok, Thanks a lot Patrick for the clarification. Your point is well noted and very much appreciated
Data selected in the years 1994, 1998, 2001, 2005 and 2007. The same questions are asked every year, but the sample is not the same. Therefore my data is not panel data but pooled cross sectional data over time.
- Badges
- Science topic