Skip to main content

StatsDunnettTest

StatsDunnettTest [/ALPH=significance /Q/Z/T=k /WSTR=wList /CIDX=controlIndex /SWN/TAIL=tailCode ] [wave1, wave2,... wave100 ]

The StatsDunnettTest operation performs the Dunnett test by comparing multiple groups to a control group. Output is to the M_DunnettTestResults wave in the current data folder or optionally to a table. StatsDunnettTest usually follows StatsANOVA1Test.

Flags

/ALPH=valSets the significance level (default val =0.05).
/CIDX=cIndexSpecifies the (zero based) index of the input wave corresponding to the control group. The default is zero (the first wave corresponds to the control group).
/QNo results printed in the history area.
/SWNCreates a text wave, T_DunnettDescriptors, containing wave names corresponding to each row of the comparison table (Save Wave Names). Use /T to append the text wave to the last column./T=k
k =0:Normal with dialog (default).
k =1:Kills with no dialog.
k =2:Disables killing.
/TAIL=tcSpecifies H0.
tc =1:One tailed test (µ<sub>c</sub> <= µ<sub>a</sub>)
tc =2:One tailed test (µ<sub>c</sub> >= µ<sub>a</sub>)
tc =4:Two tailed test (µ<sub>c</sub> = µ<sub>a</sub>) (default)
Code combinations are not allowed.
/WSTR=waveListString
Specifies a string containing a semicolon-separated list of waves that contain sample data. Use waveListString instead of listing each wave after the flags.
/ZIgnores errors. V_flag will be set to -1 for any error and to zero otherwise.

Details

StatsDunnettTest inputs are two or more 1D numeric waves (one wave for each group of samples). The input waves may contain different number of points, but they must contain two or more valid entries per wave.

For output to a table (using /T), each labelled row represents the results of the test for comparing the means of one group to the control group, and rows are ordered so that all comparisons are computed sequentially starting with the group having the smallest mean. The contents of the labeled columns are:

First:The difference between the group means
Second:SE (which is computed for possibly unequal number of points)
Third:The q statistic for the pair which may be positive or negative
Fourth:The critical q' value
Fifth:0 if the conclusion is to reject H0 or 1 to accept H0
Sixth:The p-value

V_flag will be set to -1 for any error and to zero otherwise.

See Also

Statistical Analysis, StatsTukeyTest, StatsANOVA1Test, StatsScheffeTest, StatsNPMCTest

Demos

Open Dunnett-Test Demo.pxp