We observe n individuals (consumers, firms, house-holds, etc.) at a given moment of time.
Example-, consumption expenditure on non-durables of n household in Moscow in September 2010. Often, one can assume that cross sectional data are obtained by a "true" random sampling, i.e. they are obtained as n identical, independent draws from the underlying population. For this reason, they are the easiest type of data to analyze. Cross sectional data often used in microeconomics.
We observe one variable (or several variables) over t periods.
Example, aggregate consumption on nondurables in Russia households over the period 2001-2010. In the case of time series is relevant the time aggregation of the data, i.e. monthly data, quarterly data, so t can denote the number of months or quarters, e.g. Time Series data cannot be seen as obtained as random sampling: what we observe this month clearly is affected by what we observed last month. Time Series data are more complex to analyzed and will the focus of the second part of this course. Generally, macroeconomic and financial data are time series data, though the latter are typically observed on a quite higher frequency, e.g. daily, hourly basis, while the former are observed on a monthly or quarterly base.
or Longitudinal Data.
Data sets that combine time series and cross sections are common in economics. We observe a cross section of an economic units over T period of time. For example, the published statistics of the OECD contain numerous series of economic aggregates observed yearly for many countries. Recently constructed longitudinal data sets contain observations on thousands of individuals or families, each observed at several points in time. Other empirical studies have analyzed time-series data on sets of firms, states, countries, or industries simultaneously. These data sets provide rich sources of information about the economy. Modeling in this setting, however, calls for some complex stochastic specifications.