### TABLE 5: Determinants of effort in the conditions with incomplete contracts Dependent variable: effort ICF-treatment ICR-treatment Wage .184*** (.011) .164*** (.014)

2002

Cited by 1

### TABLE VI. Logistic regression model selected using stepwise procedure

2005

Cited by 1

### Table 17). Table 17 Regression Results for SAT-10 Procedures Subscale Score and IQA Assignment Measure (R2 = .52)

2006

### Table 9: Behaviour of the forward procedure in Example 2. Selection is among all 66 regression variables.

1998

"... In PAGE 6: ...86. From Table9 we see that the best model with tree terms, containing x6, x4 and x10, only have s = 1:53 when forward regression is applied on all 66 regression variables. By using stepwise regression with FENTER = FREMOVE equal to 5.... In PAGE 6: ... In this situation there are rather small di erences in which models that are found using either FIN or R2 IN. Note however the di erence in the behaviour for these two stopping criteria from Table9 . While the values of FIN are uctuating between high and low values, R2 IN seem to be much more stable decreasing as terms are added into the model.... ..."

Cited by 3

### Table 5. Regression statistics for 2003 survey, with variables listed in order of selection by stepwise forward procedure. Significant

"... In PAGE 5: ... N requirement of wheat increases nonlinearly with grain In 2003, only three variables were selected for the yield (Ortiz-Monasterio, 2002). regression model ( Table5 ), with days between planting In 2003, on the other hand, temperatures were less and first irrigation alone explaining 18% of yield vari- favorable to wheat growth. In addition, farmers were ability.... ..."

### Table 4. Regression statistics for 2001 survey, with variables listed in order of selection by stepwise forward procedure. Significant

### Table 4: Relative Income and Cohort Size Variables Used in Analyses of Canadian Data

1997

### TABLE VII Evidence on Regulation and Political Attributes The table presents the results of running regressions for the log of the number of procedures as the dependent variable. We run seven regressions using various political indicators described on Table II and (log) GDP per capita. Robust standard errors are shown in parentheses below the coefficients.

### Table 1: Spherical distributions in ten dimensional space with two classes. Six of the variables are noise, while the other four de ne inner and outer spherical regions that de ne the classes. The results reported below the horizontal line in the table refer to our exible discriminant analysis routine using the named nonparametric regression procedure. The values are averages over 10 simulations, with the standard error of the average in parentheses. Technique Error rates Training Test

1994

"... In PAGE 26: ... We chose 250 observations from each class. The error rates of various procedures are shown in Table1 . The optimal decision boundary is the hull of a sphere with radius 3 in the rst four coordinates, and has an error rate P( 2 4 gt; 9) = 0:061099.... ..."

Cited by 79

### Table 2: Behaviour of the forward procedure in Example 1. Selection is among the ten regression variables containing the four active factors. Terms Step entered

1998

"... In PAGE 4: ... For the time being we therefore leave out of account the inert factors x5; ; x11. Table2 shows the performance of the forward selection procedure using FIN as a stopping criterion when selection is among the main e ects and the two-factor interactions for the four active factors, i.... In PAGE 5: ...Table2 also contains the values of GIN when we assume that = 0:7 is known. We observe that the values of GIN are relatively much larger for the active than for the inert terms compared to FIN.... In PAGE 5: ... We also observe that while the values of FIN increase as the rst inert terms are entered into the model, the values of GIN decrease. In Table2 we have also listed the values of R2 IN which is proportional with GIN having C = 1:21. Additional 100 responses were simulated from the same model as above, and forward selection was used to select between the ten terms containing the four active factors as before.... In PAGE 5: ... As many as 2, 4 and 5 extra terms was added for 14, 5 and 4 of the simulations, respectively. Thus, the situation presented in Table2 seems to be fairly representative for what happens in this situation. Table 3 gives the four apos;best apos; models according to residual mean square error for zero to four regression variables using best subsets regression.... In PAGE 5: ... When using FIN the model containing the four active factors are rated as number ve. However, this model also contains the four extra inert terms as displayed in Table2 . If we increase FENTER, as might seem naturally here due to the large number of terms for four factors, the true model will not be found.... ..."

Cited by 3