### Table 5: Mean Squared Errors of Forecasts: Affected Stocks

in Abstract

2007

"... In PAGE 15: ...7542 #stocks 948 494 797 information pooling. In Table5 , we report the same statistics as in Table 3, but only for the affected stocks. Across all four mergers, the bidder rm tends to produce more accurate forecasts of the affected stocks than the target rm before the merger, even though this difference is statistically signi cant only for Mergers C and D.... ..."

### Table 5: Type Equalities

2000

"... In PAGE 12: ... Soundness of (9) relies on the fact that all our recursive types are positive. It is well{known that equality of recursive types (rules (8) and (9) in Table5 ) is decidable; a simple algorithm can be found in [12]. Table 5 simply extends equality to union types.... In PAGE 12: ... It is well{known that equality of recursive types (rules (8) and (9) in Table 5) is decidable; a simple algorithm can be found in [12]. Table5 simply extends equality to union types. Since unions are nite, it is easy to verify that equality remains decidable also for our types.... ..."

Cited by 47

### Table 1: Methods for equal

"... In PAGE 44: ... Otherwise the result is that of comparing them under eq. (equal obj1 obj2) ! boolean level-0 generic The result is determined by whichever of the methods de ned in Table1 is applicable. It is implementation-... ..."

### Table 4: Type Equalities

"... In PAGE 12: ...aw (9) requires to be contractive in , i.e. occurs in only under 7!. It is well{known that equality of recursive types (rules (8) and (9) in Table4 ) is decidable; a simple algorithm can be found in [12]. Table 4 simply extends equality to union types.... In PAGE 12: ...aw (9) requires to be contractive in , i.e. occurs in only under 7!. It is well{known that equality of recursive types (rules (8) and (9) in Table 4) is decidable; a simple algorithm can be found in [12]. Table4 simply extends equality to union types. Since unions are nite, it is easy to verify that equality remains decidable also for our types.... In PAGE 17: ... A statement such as ?j P : asserts that, within the context ?, the intentions of P are those speci ed by . In the conclusion of the inference rules, we will write 1; 2 to denote the canonical form of the type 1; 2 obtained by applying the rules in Table4 . For instance, we write that from ?j P : ` 7! 7! we derive ?j out(t)@`:P : ` 7! 7! ; ` 7! fog 7! ? but actually we mean ?j out(t)@`:P : ` 7! [ fog 7! .... ..."

### Table 2. Coercion for equality=

"... In PAGE 7: ... In Lorel, a variable X can be assigned to either an atomic value, an atomic object, a complex object, or a set of objects. Table2 presents the coercion rules for equality. The coercion rules for inequality are similar.... In PAGE 7: ...quality. The coercion rules for inequality are similar. Again, the symmetric cases are not shown. Note that some of the cases in Table2 were covered in Section 3.1.... ..."

### Table 1 Equality con

"... In PAGE 19: ...trict factorial invariance with respect to groups and conditions (i.e., 4 groups) is investigated by fitting a series of increasingly restrictive models. These models as well as the restrictions imposed are presented in Table1 . In the first step, no between-group restrictions are imposed.... In PAGE 24: ...freeing the factor means of three groups (cf. Table1 ). In light of the different factor loading of the Numerical Ability subtest in the minority group, stereotype threat condition, it does not make sense to restrict this particular intercept.... In PAGE 29: ...oth conditions. For our modeling approach this poses no problem. We expected measurement bias because of stereotype threat in the female group. We again use the steps given in Table1 to assess the tenability of restrictions over these three groups. Results Except for the Math Persistence test scores of the male group in the stereotype threat condition10, the kurtosis and skewness values are in the moderate range, making the data suitable for Maximum Likelihood estimation.... In PAGE 35: ... As the four subtests were expected to load on a general arithmetic ability factor, we fitted a single common factor model in the confirmatory factor analyses. We again follow the stepwise approach given in Table1 , this time involving six groups. We expected to find measurement bias for females in the stereotype threat condition.... In PAGE 38: ...e., steps in the lower part of Table1 ) may depend on the particular parameters, which were freed in previous steps because of high modification indices. In addition, within a particular test setting one would normally test for strict factorial ... ..."