Dear Statalisters,
I have a question regarding how to accurately write down the interpretation of the results from my logistic regression model.
I calculate the following logit model:
The dependent variable bOptingOut is 1 if a firm decides not to have its financial statements audited (i.e. the firm "opts out") and 0 if it has its financial statements audited (i.e., no opting-out). The results (excerpt) are as follows:
Regarding the first independent variable bNonManagerOwners (binary variable which is 1 if there are owners who are not part of management and 0 otherwise), I formulated the following hypothesis, say, H1:
"The presence of owners who are not part of the firm's management reduces the probability that a firm performs an opting-out."
And, is it correct that H1 is a one-tailed test (because I make a statement about the sign of the coefficient)?
Based on the results, when looking at the p-value of 0.000 and the 95% confidence interval, this appears to be the case: the value is negative at a 95% confidence level. As the p-value is 0.000, this is even the case at a 99% confidence level.
Is it correct to phrase my results as follows?
"The presence of owners who are not part of the firm's management has a significant negative effect on the probability of the the firm performing an opting-out (p-value < 0.01)."
Furthermore, can I, based on the above results, "accept" hypothesis H1 that "the presence of owners who are not part of the firm's management reduces the probability that a firm performs an opting-out" - or what is the correct way of concluding on this?
I know that I can "reject the null hypothesis that bNonManagerOwners has no impact on the probability of an opting-out", but this is not really the statement I want to make.
I would appreciate your thoughts on this. It seems a bit like a trivial question, but when I was actually writing down my results, I suddenly became somewhat insecure (as I remember that this can be somewhat delicate). I want to make sure I get this right!
Kind regards,
Daniel
I have a question regarding how to accurately write down the interpretation of the results from my logistic regression model.
I calculate the following logit model:
Code:
logit bOptingOut i.bNonManagerOwners i.bNonOwnerManagers i.bExtraPpa i.bMinorities i.bFamilyOwnedStrict i.bSubsidiary i.bOnlySwissOwners lncapital cntSignatoryPower lnAge avgNoShabPubs i.bHasBranches i.industry i.firmCanton, nolog vce(cluster industry)
Code:
Logistic regression Number of obs = 217,048
Wald chi2(16) = .
Prob > chi2 = .
Log pseudolikelihood = -19762.173 Pseudo R2 = 0.2711
(Std. Err. adjusted for 18 clusters in industry)
--------------------------------------------------------------------------------------
| Robust
bOptingOut | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------------------+----------------------------------------------------------------
1.bNonManagerOwners | -.475967 .1019198 -4.67 0.000 -.6757262 -.2762079
1.bNonOwnerManagers | -.247673 .0655799 -3.78 0.000 -.3762072 -.1191387
1.bExtraPpa | -.142257 .0887571 -1.60 0.109 -.3162178 .0317038
1.bMinorities | -.4899325 .2936192 -1.67 0.095 -1.065416 .0855505
"The presence of owners who are not part of the firm's management reduces the probability that a firm performs an opting-out."
And, is it correct that H1 is a one-tailed test (because I make a statement about the sign of the coefficient)?
Based on the results, when looking at the p-value of 0.000 and the 95% confidence interval, this appears to be the case: the value is negative at a 95% confidence level. As the p-value is 0.000, this is even the case at a 99% confidence level.
Is it correct to phrase my results as follows?
"The presence of owners who are not part of the firm's management has a significant negative effect on the probability of the the firm performing an opting-out (p-value < 0.01)."
Furthermore, can I, based on the above results, "accept" hypothesis H1 that "the presence of owners who are not part of the firm's management reduces the probability that a firm performs an opting-out" - or what is the correct way of concluding on this?
I know that I can "reject the null hypothesis that bNonManagerOwners has no impact on the probability of an opting-out", but this is not really the statement I want to make.
I would appreciate your thoughts on this. It seems a bit like a trivial question, but when I was actually writing down my results, I suddenly became somewhat insecure (as I remember that this can be somewhat delicate). I want to make sure I get this right!
Kind regards,
Daniel
Comment