Problem with regression analysis

[stats123123123]

CONTENTS DELETED

The author has deleted this message.

[MaxJasper]

I do this:
 
I keep my CO2 as scale variable and assign the highest numerical value of
CO2 to bad companies...in SPSS for example if 2000 is the highest CO2 value
then if you know other companies highest CO2 can go up to 3500 then add
3500+ as a value label to CO2 numerical value and assign 3500 to each company
that did not report CO2... let them be assumed as highest CO2 producing for
now.
 
Max.



 
Dear sir/madam,
 
I am working on my master thesis and I am facing the following issue:

I have an independent variable Co2 emissions of firms and a dependent
variable Return on Assets for the same firms.
I have to perfrom a regression analysis using these variables (and some
control variables) however out of the 100 firms I want to research 12 do not
report Co2 emissions.
 
Instead of deleting them I want to somehow incorporate them into my
regression analysis as bad performers. Because these firms simply list
nothing for Co2 emissions, SPSS does not use them for regression. So I gave
these firms a 0 and created a dummy variable which identified these firms
(so I said if Co2 is 0 then dummyCo2 is 1, all others are 0). Of course this
is wrong because I distort the data by adding 12x a 0 which indicates no Co2
emissions. My question is: How can I incorporate these 12 firms into my
regression analysis as bad performers without distorting my data. What do I
have to do so that SPSS puts these firms into the regression as bad
performers?
 
Only thing I can come up with is transforming the Co2 data from continous to
ordinal (low, medium, high and inactive Co2 emissions), however, I'd rather
keep the emissions as continous interval data....so what are my options?
 
Thanks in advance for any help.

  _____