GENLIN

GENLIN is available in the Advanced Statistics option.

The GENLIN procedure fits the generalized linear model and generalized estimating equations.

Note: Equals signs (=) used in the syntax chart are required elements. All subcommands are optional.

GENLIN
    {dependent-var  [(REFERENCE = {LAST**})] [(ORDER = {ASCENDING**})]}
                                  {FIRST }             {DESCENDING }
                                  {value }             {DATA       }

    {events-var OF {trials-var}                                       }
                   {n         }

    [BY factor-list [(ORDER = {ASCENDING**})]]
                              {DESCENDING }
                              {DATA       }

    [WITH covariate-list]

[/MODEL  [effect-list]    [INTERCEPT = {YES**}]
                                       {NO   }

    [OFFSET = {variable}]    [SCALEWEIGHT = variable]
              {value   }

    [DISTRIBUTION = {BINOMIAL        }]
                    {GAMMA           }
                    {IGAUSS          }
                    {MULTINOMIAL     }
                    {NEGBIN ({1**  })}
                             {value}
                             {MLE  }
                    {NORMAL          }
                    {POISSON         }
                    {TWEEDIE (value) }

    [LINK = {CLOGLOG         }]]
            {IDENTITY        }
            {LOG             }
            {LOGC            }
            {LOGIT           }
            {NEGBIN          }
            {NLOGLOG         }
            {ODDSPOWER(value)}
            {POWER(value)    }
            {PROBIT          }
            {CUMCAUCHIT      }
            {CUMCLOGLOG      }
            {CUMLOGIT        }
            {CUMNLOGLOG      }
            {CUMPROBIT       }

[/CRITERIA  [ANALYSISTYPE = {3**} [({WALD**})]   [CHECKSEP = {20**}]
                            {1  }   {LR    }                 {n   }
                            {ALL}

    [CILEVEL = {95** }]    [CITYPE = {WALD**              }]
               {value}               {PROFILE [({1E-4**})]}
                                                {value }

    [COVB = {MODEL**}]     [HCONVERGE = {0**  } [({ABSOLUTE**})]]
            {ROBUST }                   {value}   {RELATIVE  }

    [INITIAL = {number-list [NEGBIN({1**  })]}]
                                    {value}
               {'savfile' | 'dataset'        }

    [LCONVERGE = {0**  } [({ABSOLUTE**})]]    [LIKELIHOOD = {FULL**}]
                 {value}   {RELATIVE  }                     {KERNEL}

    [MAXITERATIONS = {100**}]    [MAXSTEPHALVING = {5**}]
                    {n    }                        {n  }

    [METHOD = {FISHER ({1**})}]
                       {n  }
              {NEWTON        }

    [PCONVERGE = {1E-6**} [({ABSOLUTE**})]]
                 {value }   {RELATIVE  }

    [SCALE = {MLE**   }]
             {DEVIANCE}
             {PEARSON }
             {1**     }
             {value   }

    [SINGULAR = {1E-12**}]]
                {value  }

[/REPEATED SUBJECT = variable [* variable...]

    [WITHINSUBJECT = variable [* variable...]]

    [SORT = {YES**}]
            {NO   }

    [CORRTYPE = {INDEPENDENT**      }]
                {AR(1)              }
                {EXCHANGEABLE       }
                {FIXED (number-list)}
                {MDEPENDENT (n)     }
                {UNSTRUCTURED       }

    [ADJUSTCORR = {YES**}]
                  {NO   }

    [COVB = {ROBUST**}]
            {MODEL   }

    [HCONVERGE = {0**  }]
                 {value}

    [MAXITERATIONS = {100**}]
                     {n    }

    [PCONVERGE = {1E-6**} [({ABSOLUTE**})]]
                 {value }   {RELATIVE  }

    [UPDATECORR = {1**}]]
                  {n  }

[/EMMEANS [TABLES = factor [* factor ...]]

    [CONTROL = variable (value) [variable (value) ...]]

    [SCALE = {ORIGINAL** }]
             {TRANSFORMED}

    [COMPARE = factor [* factor ...]]

    [CONTRAST = {PAIRWISE**                }]
                {DEVIATION                 }
                {DIFFERENCE                }
                {HELMERT                   }
                {POLYNOMIAL ({1,2,3,...**})}
                             {number-list}
                {REPEATED                  }
                {SIMPLE   (value)          }

    [PADJUST = {LSD**        }]
               {BONFERRONI   }
               {SEQBONFERRONI}
               {SIDAK        }
               {SEQSIDAK     }

[/EMMEANS...]

[/MISSING CLASSMISSING = {EXCLUDE**}]
                         {INCLUDE  }

[/PRINT  [CORB] [COVB] [CPS**] [FIT**] [GEF] [HISTORY[({1**})]]
                                                       {n  }
        [LAGRANGE] [LMATRIX] [MODELINFO**] [SOLUTION**[(EXPONENTIATED)]]
        [SUMMARY**] [DESCRIPTIVES**] [WORKINGCORR] [NONE]]

[/SAVE [XBPRED[({varname             })]] [XBSTDERROR[({varname             })]]
                {rootname[:{25**   }]}                 {rootname[:{25**   }]}
                           {integer}                             {integer}
       [MEANPRED[({varname           })]]
                {rootname[:{25**   }]}
                           {integer}  
       [CIMEANPREDL[({varname             })]] CIMEANPREDU[({varname)            })]] 
                     {rootname[:{25**   }]}                 {rootname[:{25**   }]}
                                {integer}                              {integer}
       [PREDVAL[(varname)]]  [LEVERAGE[(varname)]]  [RESID[(varname)]]
       [PEARSONRESID[(varname)]] [DEVIANCERESID[(varname)]]
       [STDPEARSONRESID[(varname)]]  [STDDEVIANCERESID[(varname)]]
       [LIKELIHOODRESID[(varname)]]  [COOK[(varname)]]

/OUTFILE  {CORB = 'savfile' | 'dataset'}  {MODEL = 'file'    }
          {COVB = 'savfile' | 'dataset'}  {PARAMETER = 'file'}

** Default if the subcommand or keyword is omitted.

This command reads the active dataset and causes execution of any pending commands. See the topic Command Order for more information.

Syntax for the GENLIN command can be generated from the Generalized Linear Models Response dialog.

Release History

Release 15.0

• Command introduced.

Release 16.0

• Added multinomial and tweedie distributions; added MLE estimation option for ancillary parameter of negative binomial distribution (MODEL subcommand, DISTRIBUTION keyword). Notes related to the addition of the new distributions added throughout.
• Added cumulative Cauchit, cumulative complementary log-log, cumulative logit, cumulative negative log-log, and cumulative probit link functions (MODEL subcommand, LINK keyword).
• Added likelihood-ratio chi-square statistics as an alternative to Wald statistics (CRITERIA subcommand, ANALYSISTYPE keyword).
• Added profile likelihood confidence intervals as an alternative to Wald confidence intervals (CRITERIA subcommand, CITYPE keyword).
• Added option to specify initial value for ancillary parameter of negative binomial distribution (CRITERIA subcommand, INITIAL keyword).
• Changed default display of the likelihood function for GEEs to show the full value instead of the kernel (CRITERIA subcommand, LIKELIHOOD keyword).

Example

GENLIN mydepvar BY a b c WITH x y z
  /MODEL a b c x y z.