Inquiry on Data Analysis

#1
Greetings!

What would be the proper statistical test to use in my data. I've gathered certain type of data from 2003-2006. The illustration of the data can be seen below.

Company Year Variable 1 Variable 2 Variable 3 Dependent Variable
A 2003 5 56 25 92
2004 6 58 28 79
2005 No data
2006 4 54 24 84
B 2003 5 57 27 79
2004 8 59 29 92
2005 5 52 21 58
2006 2 54 23 66
C 2003 No data
2004 6 56 26 84
2005 7 53 28 87
2006 9 58 27 75
D 2003 8 54 29 69
2004 4 51 24 45
2005 6 60 26 84
2006 7 65 21 83
E 2003 9 58 25 79
2004 No data
2005 5 46 28 84
2006 6 58 28 90


Since my data is only 4 years (some 3 years), and they are incomplete, I decided not to do time series analysis. I was just planning to use the latest data of each company and disregard the older data and do regression analysis. However, I'm hesitant because I know that it will also add information to my analysis and we already exerted so much effort to gather them. can I just use regression with the entire data and disregard whether they are in time series format?

I would very much appreciate your kind advice on the treatment of the data.

Thank you very much and best regards.
 

BioStatMatt

TS Contributor
#2
Missing data hurts! However, there are methods for compensating for missing data in time series analysis. But you are missing some big chunks. I suppose your choice of analysis depends on what assumptions you are willing to make. If you can assume that there is no difference in these variables across time, then you can pool these data across this factor. However, be prepared to justify this action.

~Matt