Use this to save results from a parallel analysis of variance.

These results are saved when an analysis is run and must be specified before running an analysis.

- After selecting the appropriate boxes, type names for the data structures into the corresponding
**In**: fields.

## Save

The table below indicates the type of structures formed for each items.

Y-variate IDs |
Text | The list of y-variate identifiers to index the results. |

Means |
Pointer | The means for each treatment term will be saved in unnamed matrices pointed to by this pointer. The treatment levels will be in the columns of the matrix and there will be one row for each Y-variate. |

Variance-covariance matrix |
Pointer | The symmetrical variance-covariance matrix of the means will be saved in unnamed matrices pointed to by this pointer. The items of the n x (n-1)/2 variance-covariance matrix will be in the columns of the matrix and there will be one row for each Y-variate. |

Effects |
Pointer | The effects for each treatment term will be saved in unnamed matrices pointed to by this pointer. The treatment levels will be in the columns of the matrix and there will be one row for each Y-variate. |

Unit variance of effects |
Pointer | The unit variance of the stratum that each treatment term is in will be saved in unnamed variates pointed to by this pointer. The variates contain the values for each Y-variate. |

Standard errors of effects |
Pointer | The standard errors of the effects for each treatment term will be saved in unnamed matrices pointed to by this pointer. The treatment levels will be in the columns of the matrix and there will be one row for each Y-variate. |

Treatment degrees of freedom |
Pointer | The degrees of freedom for each treatment will be saved in unnamed variates pointed to by this pointer. These are stored in unnamed variates for each treatment term which are pointed to by this pointer. |

Treatment sums of squares |
Pointer | The sums of squares for each treatment will be saved in unnamed variates pointed to by this pointer. These are stored in unnamed variates for each treatment term which are pointed to by this pointer. |

Treatment mean square |
Pointer | The mean of square (i.e. the sums of squares divided by its degrees of freedom) for each treatment will be saved in unnamed variates pointed to by this pointer. |

Residual degrees of freedom |
Pointer | The residual degrees of freedom in the stratum each treatment occurs in each Y-variate. These are stored in unnamed variates for each treatment term which are pointed to by this pointer. |

Residual sums of squares |
Pointer | The residual sums of squares in the stratum each treatment occurs in from the ANOVA for each Y-variate. These are stored in unnamed variates for each treatment term which are pointed to by this pointer. |

Residual mean square |
Pointer | The residual mean square in the stratum each treatment occurs in each Y-variate. These are stored in unnamed variates for each treatment term which are pointed to by this pointer. |

Treatment variance/F ratio |
Pointer | The variance or F ratio (i.e. the treatment mean squares divided by the residual mean square) for each treatment will be saved in unnamed variates pointed to by this pointer. |

Treatment F probability |
Pointer | The probability of an F statistic being more extreme than the observed F ratio for each treatment will be saved in unnamed variates pointed to by this pointer. |

Contrast estimates |
Pointer | The contrast estimates will be saved in unnamed matrices pointed to by this pointer. The contrasts will be in the columns of the matrix and there will be one row for each Y-variate. There will be one matrix for each treatment term which has contrasts defined for it. Where not all the degrees of freedom of the treatment term are used by the contrasts, extra terms giving the deviations for each treatment level will also be saved. |

Standard errors of contrasts |
Pointer | The contrast standard errors will be saved in unnamed matrices pointed to by this pointer. The standard errors will be in the columns of the matrix and there will be one row for each Y-variate. |

Contrast t-values |
Pointer | The contrast t-values (the contrasts divided by their standard errors) will be saved in unnamed matrices pointed to by this pointer. The t-values will be in the columns of the matrix and there will be one row for each Y-variate. |

Probabilities of contrasts |
Pointer | The two sided probabilities of contrast t-values being more extreme than the observed t-values will be saved in unnamed matrices pointed to by this pointer. |

## Display in spreadsheet

Select this to display the results in a new spreadsheet window.

## Save results as variates in a single spreadsheet

The saved results will be displayed in a single spreadsheet where matrices are converted to variate columns. The variate columns that are derived from matrices will be renamed with a standard name and includes a column description that specifies the treatment term and level/contrast it originates from.

If this is not selected, then the matrices will be loaded into a number of separate spreadsheets, one for each treatment term by result type (effects, SEs etc).

## Action Icons

Clear |
Clear all fields and list boxes. | |

Help |
Open the Help topic for this dialog. |