SOME MTB COMMANDS BY ROBERTS

For additional help, send me an email note to .... Dennis Roberts

The list below assumes you have access to at least Minitab Release 10. For help on any of these, pull down the HELP
menu ... click on Contents ... and click on Session Commands. Then, scroll down the list until you reach the
command/subcommand of interest ... click on it and get additional help. Here are two examples of what you would see
if you did that ... with DESCRIBE and REGRESSION.

NOTE: The standard form for help is to show you what the command would have to be at the session window after the
MTB> prompt ... with any subcommand possibilities, followed usually with more information about the output and
possibly an example. Also, the help indicates how you would access the feature using the pull down menus and/or
dialog boxes. I am assuming with this handout that you are using the Session window and are at the MTB> prompt. I
have found this to be the most efficient way to use Minitab AFTER you get familiar with the commands.
1. DESCRIBE

Command Syntax

DESCRIBE C...C <---- Command line ... which columns to use?

BY C <---- Subcommand line ... describe by subgroups

[EXTRA! There is an undocumented subcommand DECI that will let you get varying decimal place output from
describe ... ie, DECI 4. will yield 4 places] Displays descriptive statistics for each column. The number of nonmissing
values (N), the number of missing values (N*), the mean, median, trimmed mean (TRMEAN) standard deviation,
standard error of the mean (SEMEAN), minimum, maximum, first quartile (Q1) and third quartile (Q3) are calculated.
See also:

Stat > Basic Statistics > Descriptive Statistics <---- Pull down menu way

STATS

%DESCRIBE

2. REGRESS

Command Syntax

REGRESS C on K predictors C...C <---- Y and #predictors and predictors

regr c3 2 c1 c4 ... for example

NOTE: There are MANY subcommands for regress ... RESI will store the errors or residuals, FITS will store predicted
values, and PRED will make Minitab subsitute some X value(s) into the equation and solve for Y' ... these are just a
couple.

Performs simple, polynomial, and multiple

regression using the least squares method. You

can predict future observations, and store the

diagnostics for further analysis (see Residual

Analysis and Regression Diagnostics).

To plot a fitted regression line with simple

regression see %FITLINE.

To fit a model without a constant (intercept) see

NOCONSTANT.

The command BRIEF can be used to control the

amount of output.

See also:

Stat > Regression > Regression <---- Menu approach

COMMAND/SUBCOMMAND LIST

This list is alphabetical. In general, but not always, I will say what the command does ... and give a simple version of
the command. Again ... PLEASE SEE THE HELP SCREENS FROM THE HELP MENU ... TO GET MORE
INFORMATION AND OPTIONS WITH THE COMMANDS/SUBCOMMANDS. This is particularly true for
graphics ... and relevant subcommands ... they can be a BIT tricky!

C = Command; S = Subcommand; SG = Subcommand for graphics; M = Macro

ANOVA [C] Does multifactor analysis of variance.

MTB> anova c3=c1 c2 c1*c2

AOVO [C] Does one way analysis of variance when data are in separate columns

MTB> aovo c1-c3

AREA [SG] Shades in areas under points on graphs

BOXP [C] Makes a boxplot of the data.

MTB> boxp c4

CDF [C] Cumulative distribution function ... you input baseline score and MTB gives
 you area UP TO that score. Subcommand specifies distribution.

MTB> cdf 110;

SUBC> normal 100 15.

CENTER [C] Converts column of data into z scores

MTB> cent c3, c10 <---- Makes z scores of c3 and puts in c10

CHIS [C] Does chisquare contingency analysis on 2 dimensional frequency table

MTB> chis c3 c4

CODE [C] Recodes from 1 value to another

MTB> code (3)2 c3 c4 <---- Takes all 3s in c3, makes them into

2s ... and puts new data set in c4

CONNECT [SG] Connects points to make a line

COPY [C] Copies data from one place to another

MTB> copy k1-k5 c10 <---- Takes constants k1-k5 and puts

them into column c10

CORR [C] Does Pearson correlations on columns

MTB> corr c3-c6 <---- Does all pairs of rs ... c3c4, c4c6, etc.

COVA [C] Does covariances

MTB> cova c3-c6

DATA Line for entering data for commands READ and SET

DESC [C] Prints out many descriptive stats on columns specified

MTB> desc c4-c6

%DESCRIBE [M] Does more extensive desc but, on only 1 column

DOTP [C] Makes a dotplot of columns

SUBC> same. <---- Puts c2 and c3 on same X axis

END For commands READ and SET ... tells MTB you have finished data

ERASE [C] Erases columns ... BE CAREFUL!!!!!!!!!

MTB> eras c3 c4 <---- c3 and c4 are HISTORY!

EXEC [C] Runs a macro file

MTB> exec fast' <---- Runs fast' macro

%FITLINE [M] Plots regression line

MTB> %fitline c2 c1 <---- Y in c2 and X in c1

FOOTNOTE [SG] Puts footnote on graph

SUBC> footnote Graph based on 1995 data'.

GPRO [C] Switches to professional graphics (this is default setting)

GRID [SG] Makes grid lines on graphs

SUBC> grid 1.

GSAVE [SG] Saves a graph

SUBC> gsave a:pict'. <---- Saves graph as pict' on drive a:

GSTD [C] Goes into standard (old) graphics mode

MTB> gstd

HELP [C] Gets help on requested item

MTB> help describe

HIST [C] Makes histogram

MTB> hist c5

INFO [C] Gets information on the worksheet

%INTERACT [M] Makes an interaction plot in analysis of variance

MTB> %interact c2 c3; <---- Cols for ind var codes

SUBC> responses c4. <---- Dep var data

%INTPLOT [C] Makes confidence interval plot for groups in 1 way anova

MTB> %intplot c3 c4 <---- 1st c is dep var, 2nd c is groups

INVCDF [C] Converts specified percentile rank to baseline X value

MTB> invcdf .7; <---- Input PR as a decimal!

SUBC> norm 100 15. <---- Tells MTB what distribution

LET [C] Does various arithmetic operations

MTB> let c5 = c1 + c2 + c3 <---- Adds c1-c3, puts in c5

LINE [SG] Connects points on line

SUBC> line c2 c3. <---- X comes first, then Y

%MAIN [M] Plots group means in 1 way analysis of variance

MTB> %main c5; <---- Group codes column

SUBC> responses c6. <---- Dep var data in this column

%MARGPLOT [M] Makes scatterplot with freq distributions for X and Y

MTB> %margplot c3 c4

MATRIXPLOT [C] Makes multiple scatterplots on same graph

MTB> matrixplot c3-c5

MEAN [C] Finds mean for ONE column

MTB> mean c3

MEDIAN [C] Finds median for ONE column

MTB> median c3
NAME [C] Gives names to columns

MTB> name c3='Data'

ONEWAY [C] Does one way analysis of variance

MTB> onew c3 c4; <---- Dep var data in c3, group codes in c4

SUBC> tukey. <---- Does Tukey followup test

PDF [C] Probability density function ... finds relative height at X point

MTB> pdf 89; <---- Finds rel height at X point of 89

SUBC> norm 100 15. <---- Specify what dist you are using

PLOT [C] Makes scatterplot

MTB> plot c3*c4

PRINT [C] Makes on screen list of data

MTB> print c3-c5

PROJECT [SG] Makes projection lines from plotted points on graph

RANDOM [C] Generates random data from distributions

MTB> rand 100 c3; <---- How many and put where

SUBC> inte 10:20. <---- Take from integer dist from 10 to 20

RANGE [C] Finds range for ONE column

MTB> range c10

RANK [C] Ranks data in column ... but does it backwards

MTB> rank c4, c5 <---- Uses data in c4, ranks, and puts in c5

READ [C] Allows data entry for multiple columns

MTB> read c3-c5 <---- Enter 3 values on each DATA> line

REGRESS [C] Does simple and multiple regression

MTB> regr c2 1 c1 <---- Y is c2 ... 1 predictor ... X is c1

MTB> %resplots c4, c3 <---- Res in c4 ... fits in c3

RETRIEVE [C] Calls up saved worksheet

MTB> retr pulse'

RMEAN [C] Finds means of rows and stores in new column

MTB> rmean c1-c5, c6 <---- Finds means over c1-c5, puts in c6

RMEDIAN [C] Same as above ... but for medians

MTB> rmedian c1-c5, c6

ROUND [C] Rounds values

MTB> round c1, c2 <---- Rounds values in c1, puts in c2

RSTDEV [C] Same as rmean ... but for standard deviations

MTB> rstdev c1-c5, c6

RSUM [C] Same as rmean ... but calculates sum

MTB> rsum c1-c5, c6

SAMPLE [C] Take random sample FROM an existing column

MTB> rand 20 c1, c2 <---- Takes 20 at rand from c1, puts in c2

SAVE [C] Saves worksheet AT POINT IN TIME THAT SAVE IS USED

MTB> save a:data' <---- Saves worksheet data' to a:

SET [C] Lets you enter data into ONE column

MTB> set c4

SORT [C] Sorts data from a column

MTB> sort c10, c11 <---- Sorts c10 and puts in c11

STACK [C] Stacks several columns on top of one another

MTB> stack c1-c3, c4 <---- Piles c1 c2 c3 in c4

MTB> stdev c10

STEM [C] Makes stem and leaf diagram

MTB> stem c20

STOP [C] Ends your MTB session ... BE CAREFUL!!!!

SUM [C] Sums values in a column

MTB> sum c12

SYMBOL [SG] Controls type of symbol used for plots

SUBC> symb;

SUBC> type 12. <---- Makes solid square symbols

TABLE [C] Makes table of data in columns

MTB> table c2 c3; <---- Column subdivisions here

SUBC> mean c4. <---- Puts mean from c4 in cells of table

TALLY [C] Makes basic frequency distribution ... no decimals allowed!

MTB> tally c3 c5 [Note: New REL 11 allows decimals]

TICK [SG] Controls tick marks on X and/or Y axis

SUBC> tick 1 10:30/2. <---- Puts ticks of 10,12, etc. on X axis

TINTERVAL [C] Makes t confidence interval for mean

MTB> tint c5 <---- Default value is 95%

TRANSPOSE [SG] Reverses X and Y axes in graphs

TTEST [C] Does simple t test for population mean

MTB> ttest 50 c4 <---- Assumes H(0) = 50

TUKEY [S] Does Tukey followup test for onew anova command
 SUBC> tukey. <---- See onew command

TWOSAMPLE [C] Does t test for difference in means, data in diff columns

MTB> twos c3 c4

TWOT [C] Does same as above ... but data laid out differently

MTB> twot c4, c5 <---- Dep var in c4, group codes in c5

UNSTACK [C] Opposite of STACK ... disentangles data in column

MTB> unstack c1, c10 c11; <---- Take c1 ... put in c10 c11

SUBC> subs c2. <--- using two codes values that are in c22

Return to top of page ... Top of Page ... Return to MTB Resources Page ... MTB Resources Page