Custom Tables

Analyze → Tables. When the fixed-shape tables from the other procedures are not the layout you want, these two build a table to your specification.

Custom Tables

When to use it. To pivot one or more categorical variables against others, with a cell statistic of your choosing — a generalisation of a crosstab and of a means-by-factor breakdown in one tool.

Dialog.

• Rows and (nested) Rows — the categorical variables down the side.
• Columns and (nested) Columns — the categorical variables across the top.
• Statistics — what each cell shows. Choose any of: Count, Valid N, Mean, Sum, Median, Min, Max, Std. Deviation of a measure variable, or Row / Column / Total percentage. You can show several at once (e.g. count and column percentage together).
• Compare column proportions — adds a column-proportions z-test (Bonferroni adjusted) to test whether column percentages differ.

Output. The pivot table you described, with margins computed over the margin cases, and (if requested) a Column Proportions Comparisons table.

Example. StudyZone (rows) × Sex (columns) showing the mean SystolicBP in each cell; or SmokingStatus (rows) × Hypertensive (columns) showing count and column percentage together, with a column-proportions test.

Multiple Response

When to use it. When a "pick all that apply" question is stored across several variables — either as a set of yes/no dichotomies (one variable per option) or as several variables sharing a common category coding.

Dialog. Move the variables of the response set in, and choose the set type:

• Dichotomy (default) — each variable is a yes/no flag; set the counted value that means "selected" (default 1).
• Category — the variables share a category coding; the table pools the distinct categories across them.

Output. A frequency table over the set: for each option/category, the count, the percent of responses and the percent of cases. A case counts as valid if it is non-missing on at least one variable in the set.

Example. If several columns recorded different reported symptoms as yes/no, a dichotomy multiple-response table shows how often each symptom was reported, both as a share of all responses and as a share of participants.

How this is computed

This section names which method each procedure prints — the estimator, the statistic, its degrees of freedom, and the tie, missing-value and weight rules — and links the recorded convention that explains the choice. The full index is the convention inventory; every number here is validated against independent packages per the validation suite.

• Custom Tables — counts, percentages and summary statistics over the (value, weight) pairs of the cases, so a frequency weight counts a case that many times (case weighting); the mean, SD and percentiles are the same definitions as Descriptives — n − 1 and Hazen (quantiles). The column-proportion z tests are the pooled two-proportion z with no continuity correction and no multiple-comparison adjustment (column tests); the χ² rows are Pearson's without a continuity correction (Pearson χ²).
• Multiple Response — the dichotomy and category counting modes are layout rules over the same counts, not new statistics (validation); percentages are of cases or of responses as the dialog says.

See also

• Crosstabs — for the standard two-variable contingency table with chi-square and association measures.