CELLS subcommand (CROSSTABS command)

By default, CROSSTABS displays only the number of cases in each cell of the Crosstabulation table. Use CELLS to display row, column, or total percentages, expected counts, or residuals. These are calculated separately for each Crosstabulation table or subtable.

  • CELLS specified without keywords displays cell counts plus row, column, and total percentages for each cell.
  • If CELLS is specified with keywords, CROSSTABS displays only the requested cell information.
  • Scientific notation is used for cell contents when necessary.
  • BPROP overrides PROP if both are specified.
  • If BPROP or PROP is specified without specifying COUNT or COLUMN, then the observed cell counts are included in the Crosstabulation table, with APA-style subscripts indicating the results of the column proportions tests.
  • In integer mode, the PROP and BPROP keywords on the CELLS command are ignored. If no other cell contents are requested, no table will be produced.

COUNT. Observed cell counts. This is the default if CELLS is omitted.

ROW. Row percentages. The number of cases in each cell in a row is expressed as a percentage of all cases in that row.

COLUMN. Column percentages. The number of cases in each cell in a column is expressed as a percentage of all cases in that column.

TOTAL. Two-way table total percentages. The number of cases in each cell of a subtable is expressed as a percentage of all cases in that subtable.

EXPECTED. Expected counts. Expected counts are the number of cases expected in each cell if the two variables in the subtable are statistically independent.

RESID. Residuals. Residuals are the difference between the observed and expected cell counts.

SRESID. Standardized residuals 1.

ASRESID. Adjusted standardized residuals (Haberman, 1978).

ALL. All cell information. This includes cell counts; row, column, and total percentages; expected counts; residuals; standardized residuals; adjusted standardized residuals; and pairwise comparison of column proportions using the Bonferroni correction.

NONE. No cell information. Use NONE when you want to write tables to a procedure output file without displaying them. See the topic WRITE subcommand (CROSSTABS command) for more information. This is the same as specifying NOTABLES on FORMAT.

PROP. Pairwise comparison of column proportions. Indicates which pairs of columns (for a given row) are significantly different. Significant differences (at the 0.05 level) are indicated with APA-style formatting using subscript letters. PROP is only available in general mode.

BPROP. Pairwise comparison of column proportions using the Bonferroni correction. Indicates which pairs of columns (for a given row) are significantly different, making use of the Bonferroni correction. Significant differences (at the 0.05 level) are indicated with APA-style formatting using subscript letters. BPROP is only available in general mode.

Note: The BPROP keyword takes precedence when both the PROP and BPROP keywords are specified.

Example: pairwise comparison of column proportions

CROSSTABS
  /TABLES= news BY inccat
  /FORMAT=AVALUE TABLES
  /CELLS=COLUMN BPROP
  /COUNT ROUND CELL.

The column proportions test assigns a subscript letter to the categories of the column variable. For each pair of columns, the column proportions are compared using a z test. If a pair of values is significantly different, the values have different subscript letters assigned to them.

The table in this example is a crosstabulation of survey respondents who have a newspaper subscription by the income category of the respondent, with column percentages shown as the summary statistic. The percentages in the Under $25 and $25 - $49 categories both have the subscript a so the percentages in those columns are not significantly different. However, the subscripts in the $50 - $74 and $75+ categories differ from each other as well as from the subscript for the Under $25 and $25 - $49 categories. This means that the percentages in the $50 - $74 and $75+ categories are significantly different from each other as well as from the percentages in the Under $25 and $25 - $49 categories.

1 Haberman, S. J. 1978. Analysis of qualitative data. London: Academic Press.