No announcement yet.
  • Filter
  • Time
  • Show
Clear All
new posts

  • use of difference in unbalanced panel data

    I am having unbalanced panel data of 84 countries, now the objective of my empirical investigation will be done if i convert my dependent variable into log.difference(growth rate). when i do so results for my independent and control variables turns insignificant. and when i apply difference on control and independent variable the results become significant and according to theory.
    I wonder if taking difference is appropriate with unbalanced structures? and why the difference need to be applied on rest of the variables too.