I have the attached report for a standardized test (data in the report is simulated). I'm wondering how the column Mean Scale Score Range is calculated. Do the lines indicate the limits of 95% of the scores? If so, when they recommend looking at whether the lines overlap when comparing years, is this equivalent to a t-test with an alpha of .05?

This report is from here mean scale score range

  • 1
    $\begingroup$ This may be useful: cscu.cornell.edu/news/statnews/stnews73.pdf $\endgroup$ Aug 23 '18 at 21:11
  • 1
    $\begingroup$ It doesn't look like a very competently put together report. It's not clear, but one can only presume that those are supposed to be 95% confidence intervals. And while it is true that non-overlapping confidence intervals for a specific pair of years indicates a significant difference, it is not true that overlapping confidence intervals necessarily indicates that the difference is not statistically significant. $\endgroup$
    – Denziloe
    Aug 23 '18 at 21:50
  • $\begingroup$ @Denziloe I'm checking with the test issuer to confirm whether that is indeed the 95% CI. $\endgroup$
    – Michael C
    Aug 24 '18 at 13:29
  • $\begingroup$ The publisher of the test confirmed that the lines indicate the 95% CI $\endgroup$
    – Michael C
    Aug 28 '18 at 14:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.