Field Function 



Value 



Column 



Name Format 



Card 10, . . . ,k 



What is the Jth observation number xxxxx 1-5 NITT(.T) 15 



for which intermediate output is 6- in 



desired? 



There must be at least one such card for each pass through tbe data for which 

 NIT(I) <0. 



An NITI (J) value set equal to the total number of observations or blank will 

 terminate intermediate output for that pass through the data. 

 Omit these cards if all NIT(I) > 0. 



Data Preparation 



A data transformation subroutine, titled TRNS , has been added to RISK to provide 

 the user with maximum flexibility in input data format. A TRNS subroutine must be 

 written for each use of RISK. Input to TRNS is the observation vector read according 

 to the variable format specified by control card 6. The independent variable X(l) is 

 set equal to 1 by the main program. The dependent variable, all additional indepen- 

 dent variables, and the weight for the observation must be defined within TRNS. For 

 the simplest case where no transformations are to be made, the following setup for 

 TRNS would be adequate. 



SUBROUTINE TRNS (I), NR1, NR, X, R) 

 DIMENSION D(32) , X(32) 

 DO 1 I = 2,NR 

 1 X(I) = D(I) 

 X(NR1) = D(l) 

 R = D(NR + 1) 

 RETURN 

 END 



The notation in this subroutine is defined as follows: 



D = observation vector as it is read from cards or tape. 



NR = number of parameters to be estimated. 



NR1 = subscript of dependent variable in the X vector (set equal to NR+1 by main 

 program) . 



X = vector of independent and dependent variables to be used in regression. 

 R = weight or blowup factor for each observation. 



15 



