DDBA 8307B Week 9 Crosstab Output Interpretation Example

Reading the Right Percentage: Interpreting SPSS Crosstabulation Output on Which Kind of Craft Brewery Account Reorders Within Sixty Days, How Strongly Account Type and Reordering Are Associated and Which Cells Produce the Result

Student Name

Walden University

DDBA 8307B: Quantitative Business Data Analysis Using SPSS

Instructor Name

Month Day, Year

Reading the Right Percentage: Interpreting SPSS Crosstabulation Output on Which Kind of Craft Brewery Account Reorders Within Sixty Days, How Strongly Account Type and Reordering Are Associated and Which Cells Produce the Result

The Question and the Output

A composite craft brewery in Asheville, North Carolina, sold kegs and cases to 356 new accounts in the first half of 2026: 142 bars, 118 restaurants and 96 retail shops. Its sales manager wants to know whether the kind of account relates to whether it places a second order within 60 days. SPSS Crosstabs produced a table with counts, row percentages, column percentages and total percentages, followed by a chi-square test and Cramér's V. Crosstabulations invite the wrong percentage to be read. The question asks, for each kind of account, what share reordered, so the percentages that answer it are the row percentages, computed within each account type.

The Table, With the Right Percentages

Under the table, the sentence to write is: "Retail shops reordered most often (74.0 percent), followed by bars (69.0 percent), while just over half of restaurants (52.5 percent) reordered within 60 days." The column percentages SPSS also prints answer a different question, such as what share of all reorders came from bars, which here is 42.4 percent. That figure describes the brewery's account mix, not how likely each type is to reorder, and reading it as the reorder rate is the most common mistake with this output.

Table 1

Reorder Within 60 Days by Account Type

Counts and row percentages from SPSS Crosstabs.

Is the Association Real, and How Strong?

The chi-square test of independence gives χ²(2, N = 356) = 12.42, p = .002. That means a difference this large among the three groups would be unlikely if account type and reordering were unrelated in the population of accounts like these. Cramér's V is .19; by the conventions Cohen (1988) proposed, an association of this size in a table with two degrees of freedom falls between small and medium. Field (2018) notes that chi-square depends on sample size and should be paired with a measure of strength, which Cramér's V supplies. All expected cell counts exceed 5 (the smallest is 33.7), so the test's assumption about expected frequencies is met.

Which Cells Produce the Result

A significant chi-square says the table as a whole departs from independence, not where. The adjusted standardized residuals, which SPSS prints when requested, show which cells differ most from what independence would predict. Values beyond about plus or minus 2 mark cells that depart notably.

The restaurant row drives the result: restaurants reordered much less often than independence would predict. Retail shops reordered somewhat more often. Bars sit close to what independence would predict. Agresti (2018) recommends examining standardized residuals after a significant chi-square for exactly this reason, so that conclusions name the categories responsible.

Table 2

Adjusted Standardized Residuals

Requested under Cells in SPSS Crosstabs.

How the Output Was Requested

In the SPSS Crosstabs dialog, account type was placed in rows and reorder status in columns, so that row percentages give the reorder rate within each account type. Under Statistics, chi-square and Phi and Cramér's V were selected. Under Cells, observed and expected counts, row percentages and adjusted standardized residuals were requested. Requesting expected counts makes it possible to check the test's assumption that no expected count is too small, and requesting residuals makes it possible to say which cells drive the result. Column and total percentages were left off the final table to keep readers from using the wrong ones.

What the Result Does Not Show

The result shows that account type and reordering are associated. It does not show why. Restaurants may reorder less because they rotate taps and menus seasonally, because they buy smaller quantities that last longer or because the brewery's representatives call on them less often. The data contain none of these. The 60-day window also matters: restaurants may reorder on a longer cycle, and a 90-day window might shrink the gap.

What to Tell the Sales Manager

Restaurants are the account type least likely to place a second order within 60 days. Before changing how restaurants are served, it would be worth checking their reorders over 90 days and asking a few restaurant buyers how they plan purchases.

References

Agresti, A. (2018). An introduction to categorical data analysis (3rd ed.). Wiley.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Lawrence Erlbaum Associates.

Field, A. (2018). Discovering statistics using IBM SPSS Statistics (5th ed.). SAGE Publications.