META REGRESSION

META REGRESSION is available in the Statistics Base Edition. The command must be followed by a single numeric variable that denotes the dependent effect size variable. Optionally, BY (followed by the variable names to be treated as factors) can follow the command. Optionally, WITH can be followed by the numeric variable names to be treated as covariates.

The syntax represents the meta-regression procedure.

META REGRESSION variable [ BY factor_list ] [ WITH covariate_list]

[ /CRITERIA
   [ CLASSMISSING = {EXCLUDE**} {INCLUDE} ]
   [ CILEVEL = {95**} {value} ]
   [ MAXITER = {100**} {integer} ]
   [ MAXSTEP = {5**} {integer} ]
   [ CONVERGENCE = {1E−6**} {value} ]
]

/DATA
   {WEIGHT = {variable}} {VAR = {variable}} {SE = {variable}}

[ /INFERENCE
   [ MODEL = {RANDOM**} {FIXED[(MULTIPLICATIVE)]} ]
   [ INTERCEPT = {INCLUDE**} {EXCLUDE} ]
   [ DISTRIBUTION = {T**} {NORMAL}]
   [ ESTIMATE = {REML**} {ML} {BAYES}
                {HEDGES} {HUNTER_SCHMIDT} {DERSIMONIAN_LAIRD} {SIDIK_JONKMAN}
   ]
   [ ADJUSTSE = {NONE**} {KNAPP_HARTUNG} {TRUNCATED_KNAPP_HARTUNG} ]
   ]

[ /PRINT
   [ COEFF_TEST ] [ PARAMETER ]
]

[ /SAVE
   [ FIXPRED[(var_name)] ] [ SE_FIXPRED[(var_name)] ]
   [ PRED[(var_name)] ] [ SE_PRED[(var_name)] ]
   [ CIPREDL[(var_name)] ] [ CIPREDU[(var_name)] ]
   [ BLUPS[(var_name)] ] [ SE_BLUPS[(var_name)] ]
   [ RESID[(var_name)] ] [ SE_RESID[(var_name)] ]
   [ LEVERAGE[(var_name)] ]
]

[ /BUBBLEPLOT
   [ PREDICTORS = {variable_list} ]
   [ CENTER = {FALSE**} {TRUE} ]
   [ PROPORTION = {TRUE**} {FALSE} ]
   [ FITLINE = {TRUE**} {FALSE} ]
   [ CI = {TRUE**} {FALSE} ]
   [ LABEL = {variable}[({AUTO**} {RIGHT} {LEFT} {UPPER} {BOTTOM})] ]
   [ YRANGE = {value1 value2} ]
   [ XRANGE = {value1 value2} ]
]

**Default if the subcommand 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 META REGRESSION command can be generated from the Meta Analysis Regression dialog.

Release history

META REGRESSION var_name
  /DATA SE=var_name
  /CRITERIA CILEVEL=95 CLASSMISSING=EXCLUDE MAXITER=100 MAXSTEP=100 CONVERGENCE=0.000001
  /INFERENCE MODEL=RANDOM INTERCEPT=INCLUDE ESTIMATE=REML ADJUSTSE=NONE.