### Table 2: Behavior of two descriptive statistics under various assumptions

### Table 5: Various Assumptions on the Variable Treatment Effect. (standard errors in parentheses)

"... In PAGE 26: ...11): the aggregate sub-regional unobserved group effects are correlated with the training status. In column (2) of Table5 we report the estimates of the constant treatment effect model. In contrast, with the variable treatment model, the point estimate of the mean selection bias is large and negative.... In PAGE 26: ...post participation. In Appendix D.1 we derive how this affects our regression specification. The estimation results can be found in column (3) of Table5... In PAGE 27: ... We derive in Appendix D.2 how this affects the specification of the variable treatment effects. Column (4) of Table5 reports the findings of this model. The point estimates are very different.... ..."

### Table I Maximum Degree of Crosslinkling Under Various Assumptions

### Table 2. Implications between various computational assumptions.

"... In PAGE 11: ... Finally, if security is the foremost concern, then we should choose a scheme whose proof requires the weakest security assumption. Table2 shows the implications between the various computational assumptions used to prove security of our protocols. We see that the weakest assumption is that the pairing is a one-way function.... ..."

### Table 2. Implications between various computational assumptions.

2005

"... In PAGE 18: ... Finally, if security is the foremost concern, then we should choose a scheme whose proof requires the weakest security assumption. Table2 shows the impli- cations between the various computational assumptions used to prove security of our protocols. We see that the weakest assumption is that the pairing is a one-way function.... ..."

### Table 4 Macroeconomic assumptions in various models

1997

"... In PAGE 11: ... We review th e assumptions for population and income growth. D ifferences in annual populatio n growth rates are small among the projection studies ( Table4 ). Th e outlier is OECF which assumes population growth rates of 1.... ..."

### Table 1. Location patterns maximizing the welfare function for various assumptions on the number of facilities, n, in the case of no public incentive.

### Table 4: Cost of eliminating chronic poverty under various assumptions about the policy-maker apos;s information

### Table 4: Algorithm comparisons on 2GM-2GM test data, with various prior assumptions for the SG algorithms.

"... In PAGE 30: ... The restricted \window sizes quot; for the SG3 and SG4 algorithms may be e ectively limiting the sample sizes and thus increasing the deviation of the sampled proportions from the expected proportions. This may explain the results, in Table4 , for the SG3 and SG4 algorithms with the equal prior assumption. 4.... ..."