Strict Data Sufficiency in MarketSight

Overview

Chi-squared testing of data that contains some cells with very low counts (less than 5) can possibly produce unreliable results. The biggest concern in these situations is that the Chi-squared tests will fail to detect significant relationships that actually exist (Type II error).

To avoid drawing erroneous conclusions from the results of tests with low cell counts, select the option to enforce the strict data sufficiency rule. When this option is selected, MarketSight will only run Chi-squared tests if:

  • Fewer than 20% of the cells for a Row Variable have an expected cell count less than 5; AND
  • No cells for a Row Variable have an expected cell count less than 1
Did this answer your question? Thanks for the feedback There was a problem submitting your feedback. Please try again later.

Still need help? Contact Us Contact Us