Calculates equivalent deviates for Dunnett’s simultaneous confidence interval around a control (R.W. Payne).

### Options

`METHOD` = string token |
Form of the alternative hypothesis (`twosided` , `greaterthan` , `lessthan` ); default `twos` |
---|---|

`NTREATMENTS` = scalar |
Number of treatments being compared |

`DF` = scalar |
Number of residual degrees of freedom |

`REPTREATMENTS` = scalar or variate |
Specifies the replication of the treatments |

`REPCONTROL` = scalar |
Specifies the replication of the control |

`TOLERANCE` = scalar |
Tolerance for the difference between the probability for the calculated equivalent deviate and that requested by `CIPROBABILITY` ; default 0.0001 |

### Parameters

`CIPROBABILITY` = scalars |
Specifies the probability for the confidence interval |
---|---|

`ED` = scalars |
Saves the equivalent deviate |

### Description

Dunnett’s test is useful when you want to compare several treatments with a control treatment, and use a critical value that controls the chance that any one comparison may be found significant when there are no true differences. (It is designed thus to take account of the fact that you are making multiple comparisons with the control.) The test can be preformed in Genstat using the `AMDUNNETT`

procedure, and this uses `EDDUNNETT`

to calculate the critical value (i.e. the equivalent deviate of the probability distribution).

The `METHOD`

option defines the type of interval that is formed. By default `EDDUNNETT`

assumes a two-sided interval. If you set `METHOD=lowerthan`

, it assumes a lower confidence interval, assessing the one-sided test of the null hypothesis that the treatment means are not lower than the control mean. Alternatively, you can set `METHOD=greaterthan`

, to assume an upper confidence interval, assessing the one-sided test of the null hypothesis that the treatment means are not greater than the mean of the control.

The `NTREATMENTS`

option specifies the number of treatments that are being compared with the control. The `REPTREATMENTS`

option specifies their replication, in a *scalar* if they all have the same replication, or in a *variate* if their replications differ. The `REPCONTROL`

option specifies the replication of the control, and the `DF`

option specifies the number of residual degrees of freedom.

The probability for the confidence interval is specified by the `CIPROBABILITY`

parameter, and the `ED`

parameter saves the equivalent deviate.

The equivalent deviate is estimated by an iterative process. The `TOLERANCE`

option controls the accuracy of the estimation, defining the how close the probability corresponding to the calculated equivalent deviate must be to that requested by `CIPROBABILITY`

; default 0.0001

Options: `METHOD`

, `NTREATMENTS`

, `DF`

, `REPTREATMENTS`

, `REPCONTROL`

, `TOLERANCE`

.

Parameters: `CIPROBABILITY`

, `ED`

.

### See also

Procedure: `AMDUNNETT`

.

### Example

CAPTION 'EDDUNNETT examples'; STYLE=meta EDDUNNETT [NTREATMENTS=3; DF=8] 0.95; critical PRINT critical EDDUNNETT [METHOD=greaterthan; NTREATMENTS=3; DF=8] 0.95; critical PRINT critical VARIATE [VALUES=3,2,3] trep EDDUNNETT [NTREATMENTS=3; DF=8; REPCONTROL=2; REPTREATMENTS=trep]\ 0.95; critical PRINT critical EDDUNNETT [METHOD=greaterthan; NTREATMENTS=3; DF=8; REPCONTROL=2;\ REPTREATMENTS=trep] 0.95; critical PRINT critical EDDUNNETT [METHOD=lessthan; NTREATMENTS=3; DF=8; REPCONTROL=2;\ REPTREATMENTS=trep] 0.95; critical PRINT critical