HELP ---COMMANDS--- BREAKDOWN COMPUTE CHART CORRELATE CROSSTAB CURVE CUSUM EDIT ELSE END ERASE EXECUTE EXIT GIVE HELP HISTOGRAM INPUT KOLMOGOROV LET NONLINEAR ONEWAY ONPARAM PLOT POLYNOMIAL PRINT QUERY RCORRELATE RECODE REGRESSION REMARK SET SHOW STATISTICS STEPWISE TNPARAM TTEST TWOWAY WHEN WRITE Hints: 1. Enter HELP QUERY, for example, to get help on QUERY. 2. Ordinarily STATMATE is terminated with the EXIT command. However, it may be also by terminated by a control-c character. 3. A command line may be continued onto another line by ending a line with a comma. 4. In the command descriptions, parenthesises denote the default option for a modifier. For example, TYPE : (FITTED), RESIDUAL, FORECAST indicates FITTED is the default for TYPE. 5. The variable #0 is predefined in STATMATE and represents the variable containing 1,2,3,4,... COM COMPUTE Purpose: Computes residuals, fitted, and forecast values after completion of REGRESSION, CURVE, or POLY Examples: COMPUTE XRESLT,TYPE=RESIDUAL COMPUTE SALESCAST,TYPE=FORECAST Modifiers: TYPE : (FITTED), RESIDUAL, FORECAST Hints: 1. The REGRESSION, CURVE, or POLY commands must be used to first calculate the coefficients which are used by COMPUTE. 2. COMPUTE remembers the set of independent variables (X) and the dependent variable (Y) used to compute the coefficients in the regression analysis last used. 3. Missing values in the COMPUTE variable and the dependent variable influence whether the resulting value is a missing value or a computed value. 4. Does not apply to NONLINEAR command's models. CORR CORRELATE Purpose: Computes pair-wise correlations Examples: CORRELATE URBANPOP,RURALPOP,... CORRELATE DOSAGE,TIME,ORDER=YES Modifiers: SIGNIFICANCE : (NO) , YES ORDER : (NO) , YES CROSS CROSS TABULATION Purpose: Produces a two-way crosstabulation Examples: CROSSTAB FACTORA ON FACTORB CURVE CURVE Purpose: Computes curvilinear regression on two variables Examples: CURVE Y ON X,TABLE=FIT,ANOVA,PARAM,BEST=TRES CURVE Y ON X,EQUATION=LIN,ORI,QUA,LOG Modifiers: EQUATION : (LIN), HY1, HY2, EXP, HY3, LOG, ORI, POW, QUA, SQU ALL TABLE : (FIT), (PARAM), ANOVA ALL BEST : (RSQ), TRES, ORES EDIT EDIT Purpose: Allows data to be edited Examples: EDIT YEAR,RESPONSE Hints: 1. Only the variables following EDIT may be edited with the EDIT subcommands. END END Purpose: Ends a WHEN command condition -- use HELP WHEN Examples: END ELSE ELSE Purpose: Reverses a WHEN command condition -- use HELP ELSE Examples: ELSE ERASE ERASE Purpose: Erases variables from the database Examples: ERASE SALES ERASE #12 THRU END Hints: 1. Either a single variable or a set of consecutive variables can be erased. For example, ERASE A THRU C is permitted but ERASE A,C is not. EXE EXECUTE Purpose: Allows "batch" execution of commands Examples: EXECUTE CMDFILE Hints: 1. The file named must be an ASCII file. EXIT EXIT Purpose: Exits from STATMATE to the operating system Examples: EXIT GIVE GIVE Purpose: Gives attributes (name, type, missing value) to a variable Examples: GIVE NAME #15,VAR3,VAR4 -- assign the names VAR3, VAR4 to #15 and #16 GIVE MISSING AGE,-1.0 -- interpret -1 as a missing value for AGE GIVE TYPE #10,A,N,N,A -- make #10 numeric type, #11 alpha type, etc. Hints: 1. To give either a "missing value" or type attribute to a variable the the variable must first be erased (see ERASE). HIST HISTOGRAM Purpose: Produces a histogram of a variable Examples: HISTOGRAM SALES,BARS=8,VPOS=15,RANGE=17 THRU 20 Modifiers: RANGE : min THRU max VPOSITION : number-of-lines (10) TITLE : title BARS : number of bars (10) Hints: 1. Histogram divisions are determined by RANGE and BARS 2. Use single quotes (') around title INPUT INPUT Purpose: Provides a way of entering data from a file Examples: INPUT MYFILE KEEP 2,OMIT 1,KEEP 1 INPUT DATAFILE.DAT KEEP 3 AT #22 Hints: 1. When reading in alpha data, make sure that you have used the GIVE TYPE command for any variables which are to contain alpha data. 2. Make sure the input file is an ASCII file with DATA as the very first word on the first line, and no spaces before DATA. 3. A carriage return must appear at the end of the last line of your data file. 4. OMIT and KEEP define which, and how many, fields are either omitted or kept from the input file. Enough KEEP and OMIT items must be specified to account for all fields in a record. LET LET Purpose: Performs arithmetic operations, functions, or generates a series of numbers and stores the result in a new variable Examples: LET NEWA=A+33.3 LET VOLUME=AREA*HEIGHT LET XSQUARE=X**2 LET CONTROL=X-44.0 Hints: 1. Use HELP FUN for help on LET functions. 2. Complex arithmetic statements are not allowed. An expression is composed of at most one arithmetic operator (* + - / **). 3. If a constant is involved it must appear after the operator. For example, A+3.4 not 3.4+A. 4. The reciprocal of X can be obtained by using X**-1 since 1/X is not allowed. FUN LET FUNCTIONS LET YEARLAG=LAG(YEAR,3) -- lag 3 periods LET MYSAMPLE=SAM(STATES,12) -- randomly sample 12 items LET SMOOTHED=MOV(SALES,4) -- moving average with 4 terms LET TRANSFX=SQRT(X) -- square root LET GROWTH=LOG(POPULATON) -- log base e LET SHRINK=LOG10(STEMSIZE) -- log base 10 LET WGTINC=EXP(WGT) -- exponential LET PERIOD=PRD(SALES,1,1,12) -- 1,2,3,4,5,6,7,8,9,10,11,12,1,2,... LET ROWNUM=STP(RESP,4,1,2) -- 4,4,5,5,6,6,... (4=start,1=jump,2=prd) LET ODDNUMBERS=NUM(MONTH,1,2) -- 1,3,5,7,... (1=start, 2=step) LET NORMOUT=NOR(DOSE,5.0,1.4) -- Normal random nos. (mean=5, sd=1.4) LET UNIFNUM=UNI(COST,20,44.5) -- Uniform random nos.(left=20,rght=44.5) LET RUNDOWN=DNOR(LOSS,-40.2,4) -- Normal distribution points LET DELAY=DEXP(DOSAGE,3.4) -- Exponential distribution points LET WEAR=DWEI(MACHINE,2.2,4.1) -- Weibull distribution points LET FIXLOSS=DCAU(CONT,2.4,8) -- Cauchy distribution points LET ABSOLUTE=ABS(X) -- Absolute value LET SUMIT=CUM(COST) -- Cumulative LET WHOLE=INT(WGT) -- Integer or whole number Hints: 1. The first argument of any function is a variable name. ONEWAY ONEWAY Purpose: Performs a one-way ANOVA Examples: ONEWAY DOSAGE ON TREATMENT,METHOD=LSD,ALPHA=.05 Modifiers: METHOD : (NONE), LSD, DUNCAN ALPHA : (0.05) POLY POLYNOMIAL Purpose: Perform polynomial regression Examples: POLY SPEED ON #0,TABLE=ALL Modifiers: TABLE : (PARAM), (FIT), ANOVA ALL DEGREE : (2) PRINT PRINT Purpose: List variables PRINT VAR1,VARY,SALES PRINT #2,DOSAGE Hints: 1. Prints any variable in the database 2. See SET command for hardcopy output 3. Use of the #0 variable is particularly useful here to associate a case number with the data. QUERY QUERY Purpose: Display database information on variables Examples: QUERY #3,#4,INCOME QUERY #5 THRU #15 Hints: 1. No more than 20 variables may be queried with QUERY RECODE RECODE Purpose: Changes the values of data to new values Examples: RECODE AGE,1 THRU 12 AS 1 RECODE MONTH,'1' THRU '3' AS 'Q1' Hints: 1. Recodes numeric or alphanumeric data. Useful with CROSSTABS. REGRESS REGRESSION Purpose: Performs multiple regression on several variables Examples: REGRESSION Y ON X1,X2,TABLE=FIT,PAR,ANOVA,INT=YES Modifiers: TABLE : (PARAM), (FIT), ANOVA ALL INTERCEPT: (YES), NO DURBIN : (NO), YES REM REMARK Purpose: Provides a way to document a STATMATE session Examples: REMARK YOUR WISH IS MY COMMAND SET SET Purpose: Sets a number of STATMATE parameters Examples: SET COPY=HARDCOPY SET COPY=FILE MYRESFIL.OUT SET COPY=OFF SET OUT=COSTRESID SET SEED=1532 Hints: 1. The HARDCOPY option sends output to your printer but not to the terminal. 2. The FILE option sends output to the file but not to the terminal. 3. The OUT option is used with NONLINEAR. 4. A seed value cannot exceed 30000. 5. A change in the seed value will cause the LET random number functions to generate different sequences of numbers. STAT STATISTICS Purpose: Performs simple statistical calculations Examples: STATISTICS #1,TABLE=SIMPLE Modifiers: TABLE : ALL or (SIMPLE) TTES TTEST Purpose: Performs a T-test Examples: TTEST WEIGHT ON SEX Hints: 1. The class variable (second variable) must contain only two classes. TWOWAY TWOWAY Purpose: Performs a two-way ANOVA Examples: TWOWAY WGTGAIN ON FACTA,FACTB,DESIGN=NOREPEAT Modifiers: DESIGN : (NOREPEAT), REPEAT, NESTED, UNBALANCED WHEN WHEN Purpose: Allows a subset of data for a variable to be selected for analysis Examples: WHEN #0=50 THRU 100 WHEN AGE>10 AND SEX='MALE' WHEN AGE=10 THRU 40 WHEN YEAR=1982,1984,1986 Hints: 1. Only simple expressions are allowed, i.e., one AND or OR. For example, AGE>10 AND WGT=100 THRU AND SEX='MALE' is not allowed. 2. Two successive WHENs are not allowed without an intervening END. 3. The END command terminates the condition established by a WHEN. 4. The ELSE command reverses the condition established by a WHEN. WRITE WRITE Purpose: Writes data to a file Examples: WRITE YIELD,BONDS,STOCK ON MYSTOCK Hints: 1. The output is written as an ASCII file which can be read back into STATMATE with the INPUT command. 2. By removing the first line of output to the file with an editor, the file can be used to input the data into other programs. For example, dBASEIII. TNPAR TNPARAM Purpose: Performs a two-way nonparametric ANOVA Examples: TNPAR RESPONSE ON GROUP1,GROUP2,TEST=COCHRAN Modifiers: TEST : (FRIEDMAN), COCHRAN Hints: 1. In Cochran test, all values must be zero or one. ONPAR ONPARAM Purpose: Performs a one-way nonparametric ANOVA Examples: ONPARAM XYZ ON LABCLASS Modifiers: none KOLMO KOLMOGOROV Purpose: Performs a Kolmogorov-Smirnov one- or two- sample test Examples: KOLMO AFACTOR,DIST=WEIBULL,UPARAM=1.4,SPARAM=2.0 KOMLO AFACTOR,BFACTOR Modifiers: DIST : (NORMAL), UNIFORM, WEIBULL, CAUCHY, EXPONENTIAL UPARAM : (0.0) SPARAM : (1.0) Hints: 1. normal: u is mean, s is std. dev. 2. exponential: u = displacement, s = std. dev. 3. Cauchy: u=median, u-s=1st quartile, s>0 4. uniform: u=lft end, s=rgt end 5. Weibull: u=shape, s=scale 6. Use DIST, UPARAM, and SPARAM and only one variable for the one-sample test. 7. DIST, UPARAM and SPARAM do not apply to the two-sample test. PLOT PLOT Purpose: Produces scatter diagrams Examples: PLOT Y1,Y2... ON X,TITLE='DOSAGE VS YEAR' PLOT SALES ON YEAR,HRANGE=1900 THRU 1990 PLOT ? Modifiers: HRANGE : min THRU max VRANGE : min THRU max HPOSITION : no. of spaces (50) VPOSITION : no. of lines (40) TITLE : 'title' VLABEL : 'vert-label' HLABEL : 'horiz-label' Hints: 1. Single quotes (') must appear around labels and titles. 2. PLOT ? puts STATMATE into subcommand mode. 3. PLOT alone puts STATMATE into subcommand mode. STEPWISE STEPWISE Purpose: Performs a stepwise multiple regression on several variables Examples: STEPWISE Y ON X1,X2,TABLE=FIT,PAR,ANOVA,FENTER=4.1 STEP ? Modifiers: TABLE : (PARAM), (FIT), ANOVA, SEP or ALL or NONE MAXSTEP : (20) FORCE : (0) FENTER : (4.0) FREMOVE : (3.8) METHOD : (STEPWISE), BACKWARD, FORWARD Notes: 1. STEP ? puts STATMATE into subcommand mode. 2. STEP alone puts STATMATE into subcommand mode. CHART CHART Purpose: Calculates R- and X-control chart statistics Examples: CHART WIDTH ON MACHINE,TYPE=X,CONTROL=RANGE,TITLE='GAUGE DATA' Modifiers: KCENTER : (0.0) known value KSIGMA : (1.0) known value BASIS : (SIGMA), RANGE, KNOWN TYPE : (XCHART), RCHART DISPLAY : (TABLE,PLOT), NONE HMARK : (5) HFILLER : (2) VRANGE : (DATA) min THRU max VPOSITION : no. of lines (40) TITLE : 'title' VLABEL : 'vert-label' HLABEL : 'horiz-label' Hints: 1. Use KCENTER and KSIGMA only when you want the corresponding chart values to be interpreted from the value specified, i.e., a (K)nown standard. 2. Note that the default on control limits is SIGMA. If you use TYPE=RCHART, then the default of CONTROL=SIGMA is an invalid request. In this case, you must use CONROL=RANGE or CONTROL=KNOWN. 3. CHART ? avaialble CUSUM Purpose: Produces a cumulative sum (cusum) chart Examples: CUSUM WIDTH ON MACHINE,RESET=YES,HMARK=10,TARGET=0.45 Modifiers: TARGET : (0.0) DISPLAY : (TABLE,PLOT) RESET : (NO) YES HMARK : (5) HFILLER : (2) VRANGE : (DATA) or min THRU max VPOSITION : no. of lines (40) TITLE : 'title' VLABEL : 'vert-label' HLABEL : 'horiz-label' BREAKDOWN BREAKDOWN Purpose: Produces simple statistics for groups of data Examples: BREAK HEIGHT ON SEX BREAK PRODUCTION ON COMPANY Hints: 1. Subgroups are defined by the data found in the second, group, variable. Statistics are generated for each subgroup. NONLINEAR NONLINEAR Purpose: Performs a nonlinear regression Examples: NONLINEAR MODEL=GOMPERTZ.FIL,CONV=.004,MAXITER=20,REPORT=5,TYPE=FIT Modifiers: MODEL : file name (MODEL) CONVERGE : (.001) MAXITER : (50) REPORT : (5) TYPE : (NONE), FIT, FORECAST, RESIDUAL Hints: 1. MODEL file statements must be in the order: PARAM, INIT, MODEL. For example, param alpha,beta initial 22.8, 37.44 model urbanpop=alpha+exp(beta*year) 2. Limited to at most 5 variables and 5 parameters each. 3. Terminate very slow running problems with Control-C key. 4. Use REPORT to see intermediate results. 5. The last line in the model file must be followed with a carriage return. 6. When using TYPE, the SET OUT command must be used before executing NONLINEAR. RCORR RCORRELATION Purpose: Produces rank correlation coefficients Examples: RCORR VARA ON VARB,TEST=SPEARMAN,KENDALL Modifiers: TEST : (SPEARMAN) KENDALL stop : TEST : (SPEARMAN) KENDALL stop