Reputation
11,089
Top tag
Next privilege 15,000 Rep.
Protect questions
Badges
1 14 43
Impact
~458k people reached

8h
answered Seemingly unrelated bivariate probit for endogeneity: interpretation of Rho
10h
comment Seemingly unrelated bivariate probit for endogeneity: interpretation of Rho
Doesn't biprobit usually do a likelihood ratio test that $\rho=0$ at the end, comparing the likelihood of the full bivariate model with the sum of the log likelihoods for the univariate probit models?
2d
comment Why do we need to log transform independent variable in logistic regression
This has come up before here. Also, we usually make assumptions about the error terms, not the dependent or independent variables.
2d
comment Logistic/Probit Regression if the response variable is not a probability
Wouldn't the beta regression approach exclude the zeros and the ones (unless you rescale them in some way)? Also, I am not sure that tobit is appropriate: the data are not censored, since values outside the [0, 1] interval are not feasible for proportions.
2d
answered Logistic/Probit Regression if the response variable is not a probability
Jun
25
comment Huge Standard error in calculating marginal effect
Most importantly, you want to show the estimation command that preceded margins as well.
Jun
25
revised Huge Standard error in calculating marginal effect
added 44 characters in body
Jun
24
comment Difference-in-differences model: do I have to include hierarchical terms with interaction term?
You may find this paper illuminating.
Jun
22
comment Stata nlogit and nlogitrum for the nested logit model.
You might get a better answer to this question on Statalist than here.
Jun
21
comment Does it make sense to interact two continuous variables in econometrics?
Just subtract the average of x from x itself and use that variable in the regression. Then the coefficient will give you the marginal effect of x for someone who has the average x.
Jun
19
comment Margins contrast after Tobit
It's the p>chi2 parameter
Jun
19
comment Margins contrast after Tobit
Yes, but you can also see that the p-value for the null that the difference is zero is .513.
Jun
19
comment Is eviews the only object oriented GUI based statistical package on the market?
In Stata, the estimates command can do some of this. For example, you can save estimation results along with the data, so that when you open the data, they are available. There's also a tool called Project Manager that you may find interesting.
Jun
19
revised Margins contrast after Tobit
deleted 120 characters in body
Jun
18
revised Margins contrast after Tobit
deleted 199 characters in body
Jun
18
answered Margins contrast after Tobit
Jun
18
comment Correcting for heteroscedasticity in logistic regression
I think homoskedasticity is an assumption for logits. This is because you normalize the index function coefficients by the standard deviation of the error. Thus the MLE of the parameters is biased and inconsistent if the errors are heteroskedastic, unless the likelihood function is modified to correctly take into account the precise form of heteroskedasticity. Without seeing your code, it hard to tell whether what you are doing is OK.
Jun
18
comment Does it make sense to interact two continuous variables in econometrics?
You can also rescale the continuous variables before estimation, so that the zero of $\tilde x_i = x_i-\bar x$ corresponds to average value of $x$. That way $\beta_2$ is interpretable.
Jun
18
comment Visualizing relationship between independent variable and binary response
Have you tried a lowess? If you need a CI, try a fractional polynomial as well.
Jun
17
comment Logistic regression results (coefficients) counterintuitive?
With 120 observations and 6 parameters, I would worry about finite sample bias of MLE.