Researchers collect different types of data to get conclusion about a particular problem. Data is collected according to the nature of the problem. For example, if researchers want to know about the movement of stock market index over time or currency behavior in last ten year or level of exports or imports for the last twenty years, then researchers go for collecting time series data. Time series data is type of data in which values of a particular variable is taken over a period of time, such as days, months, years etc. In simple words, in time series data, cross section or a variable remains the same, while the time changes. On the other hand, when the nature of problem is to know the income level of a particular area, then data type used in this type of problem is called cross sectional data. In cross sectional data, time remains the same, while variables of cross sections change.
What is Panel data and Panel data analysis?
A combination of both time series data and cross sectional data is used where the nature of research is to collect data of time variant and cross section variant. It means that both time and cross section change. This type of data is called panel data, which is the combination of both time series data and cross sectional data. Panel data is used where, for example, researchers want to know the capital structure of non-financial firms for the last ten years.
Panel data cover both the time dimension and cross sectional dimension of a particular problem. To get accurate results, the data used in the process should first be purified, this is called data analysis and in case of panel data, this process is called panel data analysis. Panel data analysis is important before getting into estimation techniques and problem findings, because without the analysis of data, the results is mostly inaccurate and misleading.
Hausman test – panel data analysis
In panel data analysis, different techniques are applied to data analysis, but we are going to apply Hausman test, which is used test whether fixed effect model is appropriate or random effect model. Different software packages are used to apply Hausman test, but here, we are going to use E-views, which is a very good software package for data analysis and estimation techniques. To apply Hausman test in E-views, follow the following steps
- Open E-views, make a new work file and insert your all variables.
- Make data panel by clicking on Proc then click on Structure/ Resize Current Page and fill all required fields. Now, your data in panel.
- After that click on Quick then Estimate Equation. In Specification tab, insert all your independent and dependent variable.
- Click on Panel Options tab and below that select random from Period field and click OK.
- Now, regression result screen is displayed, click on View in that window, hover over Fixed/Random Effects Testing and select Correlated Random Effect – Hausman Test.
- Now see the probability value, if the probability value is less than 0.05 then fixed effect is appropriate otherwise random effect.
To see how to apply Hausman test, click on the video below and watch complete steps.