ABSTRACT
Observations on economic variables that are made over time give time-series data. Macroeconomic data is time-series data, since it deals with aggregate relationships, and microeconomic data collected on individual households or firms is cross-section data taken from sample surveys. Analysis of microeconomic cross-section data is often concerned with very different individual units, example households with high and low incomes. If data errors are of the same relative importance in the individual household units then the absolute error will be much higher for rich than for poor households. Dummy variables can only take two values: 0 or 1. Dummy variables are often included in time-series data to allow for seasonal effects. The presence of autocorrelated error terms can be tested for by checking the value of the Durbin-Watson statistic to see whether it differs significantly from a value of 2.
