{"id":586,"date":"2018-11-12T21:43:48","date_gmt":"2018-11-13T04:43:48","guid":{"rendered":"https:\/\/blogs.ubc.ca\/datawithstata\/?page_id=586"},"modified":"2019-07-12T11:16:35","modified_gmt":"2019-07-12T18:16:35","slug":"empirical-analysis","status":"publish","type":"page","link":"https:\/\/blogs.ubc.ca\/datawithstata\/home-page\/research\/empirical-analysis\/","title":{"rendered":"Empirical Analysis"},"content":{"rendered":"

 <\/p>\n\n\n\n
\n

An empirical analysis of the data can be thought of as consisting of two parts: (1) summary statistics of the variables in the data, and (2) testing the hypotheses of a theoretical model using data.<\/p>\n

 <\/p>\n\n\n\n\n\n
\n

Data Analysis<\/a><\/strong><\/span><\/h3>\n

Datasets in the social sciences usually come from secondary sources unless the researcher herself conducts an experiment or collects primary survey data from the field.\u00a0 Secondary data typically needs to be arranged in the desired format before the analysis. This is known as data management<\/a> and is very crucial for efficient empirical exercise. Once the data has been arranged properly, different properties of the data like the mean, median, mode, range, standard deviation, skewness, correlation, etc. of the variables are summarized using different graphs<\/a> and tables<\/a>.<\/p>\n

\n

This summary of the data is sometimes helpful in discovering new trends and relations among variables that were previously unexplored.\u00a0\u00a0<\/em><\/p>\n<\/blockquote>\n<\/td>\n<\/tr>\n

\n
    \n
  • \n
      \n
    • \n
        \n
      • Histogram<\/li>\n
      • Scatter Plot<\/li>\n
      • Bar Graphs<\/li>\n
      • Summary Stats<\/li>\n
      • Correlations<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/td>\n<\/tr>\n
\n\n\n\n\n
\n

Regression<\/strong><\/a><\/span><\/h3>\n

\u00a0<\/strong>Regression is a statistical tool to examine the relation between two variables in an efficient manner. The choice of a regression model or technique depends on the variable type; variables can be qualitative or quantitative; continuous or discrete, etc,\u00a0 While statistical measures like the correlation coefficient can only measure the linear relationships among variables, regressions are flexible in accommodating non-linear relations among variables. Regressions can also be used to argue about causality, and not just correlation if they are properly designed. We are going to discuss these issues in more details <\/span>here: Regression Models\u00a0<\/a><\/td>\n<\/tr>\n

\n