T-test

In a paired t-test we were comparing vowel lengths of words spoken by the same person so we could pair the lengths spoken by the same person. In this test the question we ask is whether gender affects formant frequency and we can't measure one person pronounce a word as a male and then as a female so we can't use a paired t-test.

First, open this web page. Make sure that the unpaired t-test radio button and the enter or paste up to 2000 rows radio buttons are clicked.

Second, go back to the spreadsheet and sort the data by the first column (gender). Highlight all of the F1 frequencies of the words ending in /d/ spoken by a female and copy them into the group 1 values box in the web page.

Third, copy all of the F1 frequencies of the words ending in /d/ spoken by males and copy them into the group 2 values area in the web page.

Fourth, click the calculate now button and look at the results.

On your homework report whether females have significantly higher or lower F1s than males. What do the statistics tell you?

Fifth, repeat the above steps modifying them in order to find out whether men or women have significantly different F2s on the final /d/ words.