what kind of analysis to use? regression, MANOVA, panel data analysis???

[jurgen57]

Hi,
Im an undergrad student doing some research on the effects of financial crisis of 2007 2008 on firms.
I have data from 2005 to 2010 for over 2000 firms. I have performed t tests on the leverage ratios and noticed they have changed. Now, I would like to understand where this change comes from by using variables such as size of firms, profitability, industry, inflation, gdp etc. I do not know what kind of analysis to use thought, because most of these dependent variables have 2 values, pre crisis and post-crisis.

Any help is appreciated.
Thank you

[noetsi]

This won't make your life any easier, but the problem with regression or Manova is that they are essentionally cross sectional. They don't assume change over time. And they assume that data points are independent, which won't be true normally in time series data.

I have been looking for a year for the question you ask. If I find one I will let you know. Some form of ARIMA combined with covariates might apply, but I am not sure.

[spunky]

i am a little bit confused here... are you interested in looking at changes in trend over time with particular emphasis on what happened before and after the crisis? you mentioned that you have data from 2005 to 2010 but then you said that you only have pre- and post- crisis information. depending on what you have, you can either try a mixed effects model (which i think would be desirable because at 2000 firms with data spanning 5 years it sounds like a very interesting, juicy dataset to play around with) or you can do this through analysis of difference scores (which is not as interesting as linear mixed effects models but it is very helpful when you only have data on two timepoints).

i guess i'll wait to see what kind of data you have befor i elaborate further....