Skip to main content

Chapter 7

Troubleshooting

When something does not work as expected, first try to reproduce it with the smallest safe example:

All chapters Download PDF

Troubleshooting

Start With A Small Reproduction

When something does not work as expected, first try to reproduce it with the smallest safe example:

  1. Save the project.
  2. Try the same action again.
  3. Try the workflow with built-in sample data.
  4. If the issue only appears in your data, make a small synthetic dataset with the same column types and missing-data pattern.
  5. Record the exact menu path, selected variables, options, warnings, and error message.

This approach separates app behavior from dataset-specific problems and makes support requests safer.

Import Problems

If a file does not import correctly:

  • Confirm the file is a plain CSV when using CSV import.
  • Check whether the first row contains column names.
  • Convert legacy encodings to UTF-8.
  • Remove report titles, footnotes, merged cells, blank trailing columns, or subtotal rows.
  • Open the file in a text editor to confirm delimiters are consistent.
  • Check whether decimal separators, dates, or quoted text use a format SCCalc can infer reliably.

After import, immediately review Data Editor and Variable View. Most import problems are easier to fix before transformations and analyses have been run.

Permission Errors

SCCalc is sandboxed. If opening, saving, or exporting fails:

  1. Retry the operation.
  2. Select the file or destination folder through the macOS open or save panel.
  3. Avoid moving project files while SCCalc is open.
  4. Save a new copy if the original location no longer grants access.
  5. Export to Documents or Desktop to test whether the issue is folder-specific.

Do not edit app bundle contents or generated manual files directly. Use normal project, import, and export commands so macOS can manage permissions.

Unexpected Analysis Output

If results look wrong:

  • Confirm variables have the expected measurement levels and numeric types.
  • Check missing data and filtering choices.
  • Check whether split file or weighting is active.
  • Inspect outliers and impossible values.
  • Verify grouping variables contain the intended categories.
  • Confirm reverse-coded items and scales were prepared correctly.
  • Re-run the analysis from saved data to confirm the result is reproducible.

If the same procedure works with sample data but not your dataset, inspect the specific variables used by the procedure. Look for imported text, sparse groups, empty variables, invalid values, and labels that differ by spelling or case.

Export Issues

If a PDF, chart, table, or report export fails:

  • Choose a different folder such as Documents or Desktop through the save panel.
  • Use a new filename.
  • Close preview apps that may be holding the file open.
  • Confirm there is enough disk space.
  • Export a smaller result or a different format to isolate the issue.

Before sharing a failed export for review, open the file and remove or recreate any confidential data, labels, comments, paths, or results. If the file cannot be sanitized, describe the issue using sample data instead.

Online Help Not Available

If online resources do not open:

  • Confirm the Mac has network access.
  • Use Help -> User Manual for the bundled offline PDF.
  • Use Help -> SCCalc Help for the in-app help center.
  • Use Help -> SCL Quick Reference for local command syntax guidance.
  • Try the online manual later if the issue appears to be network-related.

The offline manual may not include the newest website corrections, but it is the best source when working without a network connection.

When To Contact Support

Contact support when a workflow fails repeatedly with the same data and options or when a result cannot be explained by data, metadata, filters, weights, or procedure settings.

Include:

  • SCCalc version.
  • macOS version.
  • Import format and approximate dataset size.
  • Analysis procedure and selected options.
  • Any warnings shown in Results.
  • Whether the same issue appears with sample data.
  • A sanitized or synthetic reproduction when possible.

Do not send raw confidential data, unredacted project files, private screenshots, passwords, tokens, or secrets. If support needs an example, provide sample data or a small synthetic dataset that reproduces the issue without exposing real people, organizations, research subjects, customers, or unpublished results.