### Table 6 Estimated Spectral Densities of Sampled Parameters,Tobit Censored Regression Model

1992

"... In PAGE 22: ...mode the 118-dimensional density relevant for Gibbs sampling with data augmentation is high (by definition), and given the smoothness inherent in this problem movement in the various dimensions is more likely and hence convergence is more rapid starting from this point. (4) There are quite substantial differences in the serial correlation properties of the sampled parameters, as indicated in Table6 . The parameters b 1 and b 3 exhibit very strong positive serial correlation, to the extent that a poor picture of the pattern emerges with only 400 observations.... ..."

Cited by 105

### Table 8: Output Loss Regressions

2004

"... In PAGE 19: ...ssociated with output gain. We borrow this variable from Frankel and Wei (2004).40 Given the nature of the dependant variable, which is censored to the left of the distribution at zero, the natural candidate for the estimation is a Tobit model. We perform Tobit and iv Tobit regressions and report the typical results in the first two columns in Table8 in Appendix A.5.... In PAGE 19: ... Finally, the Lag of reserves in months of imports is typically negative and statistically significant, meaning that the presence of a large stockpile of reserves tends to dampen the recession that might come in the aftermath of an external crisis. In columns 3 and 4 of Table8 , we also report IV linear and IV probit results which depict a similar picture. The only noticeable difference is that in the IV probit regressions the coefficient on openness is typically statistically significant.... ..."

### TABLE 3 Regression Model

in Public Utility Companies: Institutional Ownership And The Share Price Response To New Equity Issues

### Table 1. Regression Models

2006

"... In PAGE 7: ...Table1 is straightforward. Because of the lack of space we will only present the interpretation for Assistance_Regression_Model.... In PAGE 7: ...or only 2.28 points for every hint request averaged over all questions. 3.4 Tracking Two Outcomes Longitudinally Given Table1 , we constructed the following formulas to compute values for the two new variables that represent student knowledge in a certain month: original_predicted_score = 4.... In PAGE 8: ... The non-linear model probably over-fitted the training data and was thus dis- regarded. In both regression models presented in Table1 , pretest was a significant parameter. We wondered how much the tutoring and assistance information can help without pretest because pretest scores are not always available every school year.... ..."

Cited by 9

### Table 1. Regression Models

2006

"... In PAGE 7: ...Table1 is straightforward. Because of the lack of space we will only present the interpretation for Assistance_Regression_Model.... In PAGE 7: ...or only 2.28 points for every hint request averaged over all questions. 3.4 Tracking Two Outcomes Longitudinally Given Table1 , we constructed the following formulas to compute values for the two new variables that represent student knowledge in a certain month: original_predicted_score = 4.... In PAGE 8: ... The non-linear model probably over-fitted the training data and was thus dis- regarded. In both regression models presented in Table1 , pretest was a significant parameter. We wondered how much the tutoring and assistance information can help without pretest because pretest scores are not always available every school year.... ..."

Cited by 9

### Table 2. Regression Models

2006

"... In PAGE 6: ...f-F-enter lt;= 0.05 and Probability-of-F-remove = 0.1 to predict MCAS scores. In Table2 , we summarize the five models that have been built for which we selected different groups of independent variables (IV) for regression. For all these models, we saw normally distributed standardized residuals.... ..."

Cited by 6

### Table 3. Regression Results

"... In PAGE 17: ... Year dummies, D1997 and D1998, are included in the model to control for systematic differences across time. A Fixed-Effect model is estimated and reported in Table3 assuming that country- specific effects vary systematically among the exporting countries. The results for alternative specifications are also presented in the table.... ..."

### Table 3: Regression Model

"... In PAGE 13: ... A quadratic model is used within a data- fitting problem, with constraints being set on the eigen-decomposition of the quadratic term. Table3 shows the results for one particular sets of constraints.... In PAGE 13: ... As well as the improved run time obtained with MAD we also note that the error in the position of the minimum obtained was 2 orders of magnitude smaller than that obtained using finite-differencing. Interestingly, the bottom row of Table3 shows the mode of AD for the constraints shifting automatically from full, to sparse to compressed as the problem size increases. 5 Conclusions and Future Developments In this paper we have presented the MADJacInternal function which enables auto- mated, performance driven selection of a Jacobian evaluation algorithm via the forward mode fmad class of the MAD package.... ..."

Cited by 2