# AM I allowed to do this?

P

##### Guest
Basically i have a binary variable called DP ( takes values 0 or 1) that i had to drop from my multiple regression due to few observations with of the value 0 ( hence statistically NOT representative). My study is concerned with firms in two countries. Basically even though i have to drop the variable DP, i wondered if i could write just a quick something about its affect on the dependant variable ROA.The following is what i want to know if im allowed to write, and any comments would be appreciated and thanked;

START OF EXTRACT

The variables DP = 0 (dividend not paid) and DP = 1 (dividend paid) were excluded from the rest of the analysis because only 1 UAE company and 5 Saudi companies paid dividends, and therefore the sample sizes of the companies which did not pay dividends were not sufficiently representative to perform statistical analysis on the variable DP.

However although it cannot be established as a definitive result, it can be inferred (from table 3) that the ROA for firms in both countries is consistently higher when a company has a policy to pay dividends (i.e. DP = 1)

Table 2 Mean & Standard Deviation of DP

USA
MEAN STANDARD DEVIATION
DP = 1 0.082123 0.07453
DP =0 0.011438 0.002968

UK
MEAN STANDARD DEVIATION
DP = 1 0.10341 0.07453
DP =0 0.036554 0.002968

END OF EXTRACT

#### bugman

##### Super Moderator
Sounds reasonable, though I think there are log linear models availble for small sample sizes (though maybe not this small).

A suggestion about the syntax.. you wrote:

However although it cannot be established as a definitive result...

NOTE

A result is a result regardless. So to say it is cannot be established as a definitive result is incorrect.

How about: However, although no firm conclusion can be drawn from these results, it can be inferred from the limited available data (Table 3) that ....