### Table 1. Separation errors of the tested robust methods.

### Table 4 Overall Fit, Robustness and Model Comparison

2002

"... In PAGE 36: ... These correlations are merely indicative, as the number of means (4 or 6) is too few to obtain reliable differences between correlations. Therefore, we also analyzed the interaction or main F-test of interest for each experiment (see Table4 for a list of these effects). Although we also conducted t-tests to make direct comparisons, these are not reported as long as the results are identical to those mentioned earlier for the specific paradigms.... In PAGE 36: ... To determine a standard of comparison for the robustness analyses, we first assessed the overall fit of the feedforward simulations previously discussed. As can be seen in the top panel of Table4 , the feedforward simulations resulted in very high correlations ranging from .89 to 1.... In PAGE 36: ...eduction, the mean correlation decreased by -.07 (mean r = .87). As can be seen in Table4 , in most cases the decrease was marginal and substantial only for the prohibition experiment (Simulation 1). The F-tests of interest remained significant in all simulations and their pattern did not change substantially, except in the prohibition simulation where the interaction showed a crossover that did not appear in the original experiment.... In PAGE 38: ...1) that provided the best fit with the data. As can be seen in Table4 (bottom panel), Shultz and Lepper apos;s (1996) consonance simulations achieved very high correlations with the human data, which ranged from .... ..."

Cited by 8

### Table 5. Verifying Electricity Effect: aggregate data

"... In PAGE 15: ... To empirically separate the electricity effect from income, we use data at the national level where we have all three variables concerned: income, asset index, and access to electricity. Table5 summarizes our results for verifying electricity effect. The estimated coefficient for access to electricity is hardly affected by adding income variable in the model specification, after controlling ether only for asset index (mod-1 vs.... ..."

### Table 1. Relative predictive performance of the robust and non-robust estimators. Mean of 25 Model Errors (G:

"... In PAGE 7: ... Four underlying functions used in the Monte Carlo experiment. We use the four test functions de ned in Table1 of Donoho and Johnstone1: blocks, bumps, heavisine and Doppler. We normalize them such that their \standard deviation quot; is equal to 7, Z 1 0 (f(x) ? f)2dx = 49; where f = Z 1 0 f(x)dx: The four functions are plotted in Figure 2.... In PAGE 8: ...Based on Table1 , we see that the simulation results can be separated into two groups: the predictive performance of the robust procedure is poor for bumps and blocks, and good for heavisine and Doppler. This comes of no surprise... ..."

### Table I. Robust standard errors are shown in parentheses.

2006

Cited by 1

### Table 7. Sample Splits for Multinationals versus Other Firms Defining MNCs as gt;= 5 %

2004

"... In PAGE 28: ... The Worldscope database allows us to identify which firms are multinationals, as defined as firms with at least 5 percent total assets owned abroad. In Table7 , we redo the basic specification, but we separate firms with foreign assets from other firms. In columns (1) and (2) we separate firms with foreign assets by defining multinationals as firms with at least 5 percent of total assets as foreign.... In PAGE 29: ... However, the Ivory Coast is an unusual case. The results in Table7 appear to indicate that domestic firms overall (whether joint ventures or otherwise) benefit from incoming FDI through a relaxation of financing constraints. 6.... ..."

Cited by 1

### Table 5: Robustness Check for NEWENT Regression

"... In PAGE 13: ... Given the first two categories, we worked with four measures of both GROWTH and NEWENT in each stage of our analysis: the baseline and alternative measures in both the full data set and the sub- sample obtained by excluding the outlier regions. OLS estimates of (1) obtained using the four measures of NEWENT are presented in Table5 . Several comparisons with the estimates in Table 4 (reprinted in the first row of Table 5) are notable.... In PAGE 13: ... OLS estimates of (1) obtained using the four measures of NEWENT are presented in Table 5. Several comparisons with the estimates in Table 4 (reprinted in the first row of Table5 ) are notable. First, the statistical significance of TPRIV, LPRIV and CONTROL is robust to the exclusion of the outlier regions, but the quantitative significance of each... ..."

### Table 1. Quantitative evaluation of robustness in terms of the hpf range.

2000

"... In PAGE 8: ...nd CSF. Only in two cases, parts of the superior sagittal sinus have been included. The results were satisfactory even where the range of proper pre-flooding heights hpf has been rather narrow. This has been the case in six of 90 simulated images due to a slice thickness of 9 mm, which is too thick for a good separation of eyes and frontal lobes (compare Figure 5 right), and in two of 43 clinical datasets (see Table1 ). When inspecting these two images, we found that some isolated bright voxels existed in ... ..."

Cited by 9

### Table 1: Results of the OLS regression with robust standard errors

"... In PAGE 6: ... Thus, the models satisfy the conditions for giving unbiased, consistent and efficient estimates of regression coefficients. Results of the OLS regression analysis with robust standard errors are presented in Table1 .The results support our first hypothesis that SOA adoption improves electronic supply chain effectiveness.... ..."

### Table 7 | Explaining Agreements (robust Tobit estimates)

"... In PAGE 16: ... This clearly supports our hypothesis. The robust Tobit estimates that are reported in Table7 of the Appendix strengthen these findings. There the coefficients of W FAIR and L FAIR are both positively significant at the 1 percent level (two-sided tests).... In PAGE 30: ...In the main text we found with the help of Spearman rank order statistics that the fairness jugdments of losers and winners are significantly positively correlated with the agreement (in winner share) reached in a bargaining pair. The regression results shown in Table7 corroborate this finding.23 The fairness judgment of losers as well as the fairness judgment of winners exhibit a highly significantly positive coefficient (p lt; 0:01 for both variables, one-sided tests).... ..."