Given an F-Statistic, is there a way to find the P-Value with a TI-89?

March 5, 2015 4:13 PM

One of my Stats II homework questions is: "Five groups are being compared, with 13 observations per group. The value of the F statistic is 1.61. Obtain the P-value of the test." I found an online calculator but won't have access to that during tests. Is there a way to find this on the TI-89?

The examples in the text give all the info regarding groups, means, etc. I think I can do the ANOVA given all that info but I can't see a way to just find the P-value itself.

The possible answers are:
0.00063
0.00253
0.0431
0.1835
0.2341

I looked at the tables in the text and found the critical values for 4 and 60 degrees of freedom:
alpha=0.05 2.53
alpha=0.10 2.04

So if the F-statistic is 1.61, the P-value is greater than 0.10 and I can rule out the first three answers. I couldn't find a table with alpha=0.2 so I can't figure out if it's 0.1835 or 0.2341. If there isn't a way to find this on the TI-89, can someone point me in the right direction for solving this? I feel like I'm missing something obvious here. I was doing great in the class until we started this chapter. Thank you for your help!

posted by Beti to Education (5 answers total)

Best answer: Yes. Input the formula for your F-distribution (with appropriate parameters d1 and d2) into your graphing utility then integrate from your observed value -> infinity.
posted by dilaudid at 4:18 PM on March 5, 2015

Although- is there an easy way to use the beta function on that calculator? Sorry- my previous answer may be useless.
posted by dilaudid at 4:25 PM on March 5, 2015

Response by poster: Oh, excellent! Thank you so much. I knew there must be some way to do this. I'm not sure what a beta function is. But your first answer works. Thanks again!
posted by Beti at 4:26 PM on March 5, 2015

You can also look up the inverse CDF directly. The instructions here look right, although I don't have a TI-89 handy to check. You're dealing with the first type of problem. To find the p-value, you'd want the "Lower Value" to be the statistic and the "Upper Value" to be infinity.
posted by matildatakesovertheworld at 9:43 PM on March 5, 2015

While this solution works, I can't imagine it to be what your lecturer wants you to do, unless they are deeply sadistic. I would ask them about their intention with regards to calculating the p value here, it seems unlikely that they want everyone to do integration on their calculator under exam conditions!
posted by Cannon Fodder at 12:09 AM on March 6, 2015 [1 favorite]