Features

[embar]

Thanks for useful software. In Vilnius University (Lithuania) we use SPSS,
and most important functions exist in PSPPIRE. But  there are some
fetures, that is missing. So  I made a list of features, that we need,
that is realised in PSPP 0.7.8-ge98c31; legend:
+ found in PSPPIRE, required
- not found in PSPPIRE, required
* found in translation file, not found in PSPPIRE, required
~ partly found, required
? not found, not required

Common:
~ Select cases (missing option filter by condition)
+ Weight

Graphs:
~ (seems realised for Descriptive)

Transform:
~ Compute (missing option "If..." for condition)
+ Automatic recode
+ Recode

Descriptive statistic:
+ Frequencies
+ Descriptive (mean, min, max, z-values...)
+ Explore
~ Crosstabs (+ main dialog;
             - missing option McNemar in [Statistic...],
             ? dialog for [Cells] may be improved)

Compare Means:
- Means (- [Options]: - ANOVA table and eta)
+ One-sample T test
+ Independent-Samples t test
+ Paired-samples T test
~ One-wayANOVA (~ [Contrasts]: - Trend
                * [Post Hoc] (*Bonferroni, *Tukey)
                + [Options]: +Descriptive, + Homogeneity of var
                             - Means plot
                             ? Other options will not be used in our studies)

Correlate:
~ Bivariate (+ Pearson;
             - Spearman
             ? Kendall tau-b (missing, but will not be used in our studies))

Nonparametric Tests:
+ Chi-Square
- Runs
- 1-sample K-S
+ 2 related samples (+ McNemar, Sign, +Wilcoxon
                     ? Marginal Homogeneity missing, but will not used in
our studies)
+ K related (+Friedman)
- 2 independent samples (- Mann-Whitney;
                         ? K-S Z, W-W runs, MER will not used in our studies)
- K independent (- Kruskal-Wallis H,
                 ? Median, J-T (will not be used in our studies))

General Linear Model:
- Univariate
- Repeated measures

Regresion:
~ Linear ([Statistics]: - Durbin-Watson, - collinearity, - part and
partial correlations
          [Save]: Cook's distances)
- Curve estimation
- Binary logistic

Scale:
+ Reliability analysis


Classify:
- Discriminat
- Hierarchical cluster
* K-means cluster

Data reduction:
~ Factor (- KMO and Bartlett)

Having all these features may fully replace SPSS.

And GUI of PSPPIRE is not friendly for manually entering data: you must
manually add variables before entering data; you must manually add cases
if cases not fit to window, copy/pasting don't add variables by
necesarity.

Mindaugas B.