Data in which changes are introduced naturally, rather than by a researcher. For example, governments in different countries have different levels of spending. This would allow for an analysis of the effect of government spending on growth. (Not surprisingly, this has been studied, and increased government spending is closely associated with reduced economic growth). Forecasters often rely on quasi-experimental data. In contrast to <a href="search.pl?p=experimental data">experimental data</a>, there are many threats to validity, so <a href="search.pl?p=eclectic research">eclectic research</a> might be useful.<o:p></o:p></P>