Quote:
Originally Posted by jatx
But riddle me this:
Which regressors best explain the level of economic development in a country? How are you measuring the level of economic development? Are you working with a metric which is also closely correlated with social stability? Which is the cause and which is the effect?
|
To measure economic development and other factors I used a variety of factors: GDP per capita, income disparity and unemployment rates. My hypothesis was that these factors were closely tied with social stability and could be influencing factors on intrastate conflict.
An interesting question of cause and effect-- Nations which host intrastate conflict often DO have economic problems, and one may be led to believe that economic influences do impact conflict, but this relationship could be spurious-- the dependent may actually be the independent. That is to say intrastate conflict causes economic problems, another hyothesis which should be tested.
Quote:
Originally Posted by jatx
Unless you (1) have all of these factors straight in your mind and (2) your objective function accurately reflects this understanding, you will have multicollinearity. Your error terms will be huge and your R-squared will be minuscule.
|
Because the relationship between independent and dependent variables is possibly spurious, the time order is incorrect, could b explain of my lack of statistical significance. My data analysis actually suggest a negative correlation but with a high p-value was labeled unreliable results.
Overall I found that economic factors accounted for less that 4-8% of variation of intrastate conflict- a very low R-square. As you suggested these variables are multicollinear and it is difficult to determine the impact of any single variable.
Rm1249