TI83/84: F Critical (Inverse F) using Solver and Fcdf

[@Damien-dl2dz]

Thanks a looootttt! I was upset and even gpt-4 didn't think to that solution, you're the boss!

[@SBarton720]

Thank you so much! I wish I found this video sooner.

[@aaronhyman4764]

Hi, I have an exam on this stuff tomorrow. How do you do this for a 2 tail test? Thanks

[@quantquill]

Hi Aaron, split the alpha in half for a two tailed test. That's the only difference in getting the F critical value.

[@aaronhyman4764]

@@quantquill For upper-tail critical value of a 2 tail test, level of significance of 0.02, n1=10, n2= 21 it would look like this in my calculator? Lower: X Upper: e99 dfNumer: 20 dfDenom: 9 (Function) - 0.01 ?? = 4.81 but apparently that is wrong

[@quantquill]

@@aaronhyman4764 Hi, Aaron, you're getting exactly the correct critical value based on the numbers you've entered, but what you want to do is double check that you're using the right D1 and D2. In an F test, we make the largest sample's variance our variable 1, and therefore it's sample size - 1 = our D1. So if your sample with n = 10 has the larger variance, then D1 = 9. The other sample size, which has the smaller variance, would be used to get D2 = 20. In that case, you'll get F critical of 2.984. You can also double check that your numbers in the calculator are correct if you have access to Excel or Google Sheets using =f.inv.rt(probability,D1,D2). Prob. on the two-tailed would be alpha/2, or 0.01 in this case. When I check them using your D1 of 20 and D2 of 9, I also get 4.808, which is verification that the calculator is returning the output correctly. Double check your sample variance to see if you need to reverse your D1 and D2.

[@BrownPen_BluePen]

are you able to make an program or use equation solver for the pearson critical value? my teacher told us its impossible and to use the chart for r.

[@quantquill]

Most people do use the table because it's easy to look up. However, you can calculate it using the formula r critical = sqrt((t^2)/(t^2 + df)), where t is the t critical value given alpha and df = n - 2. Example: We are running a one-tailed test with alpha = .05, and n = 102 sets of data. Therefore, df = 100. We can use invT in the calculator to get the t critical value. Plug in 0.05 for alpha and 100 for df. (If you're running a two-tailed test, chop alpha in half, but let's say this is one-tailed). With alpha = .05 and a one-tailed test, you'll get t critical = -1.660234. We don't care about the minus sign, we're going to square the t critical in the formula anyway. Next step: solve for r critical. sqrt(1.660234^2/(1.660234^2 + 100)) = .163782, rounds to 0.164. This is the same value we get when we look up Pearson's r critical value for a one-tailed test with alpha = .05 and df = 100 in the table. One could set up a program in the calculator to do this two-step calculation for us using similar tactics to how the marvelous IntroStats set up the invT program (kzbin.info/www/bejne/a3fXZpiQi6-nhs0). Once the calculator knows what you want to use for alpha and df, it would solve for inverse t, plug that t value and df into the square rooted formula and spit out the Pearson correlation critical value. (I'm not planning to write this program, since I'm not asking my students to look up this critical value.) If someone takes on the challenge, or knows of someone who already wrote such a program, please post a link here. Hope this helps! And please let me know if I've misunderstood your question.

[@Karymald]

Mines had e2 and I can’t seem to solve it

[@ccvv568]

On the equation solver, my calculator shows "bound" how do you clear it out and change to eqn?

[@quantquill]

Hi, Jin, you should be able to click the up arrow to put your cursor next to that and click the Clear button.

[@blaizakin90]

no matter what i put it puts out -10 for x

[@quantquill]

I'm sorry you're having this issue. Not sure what would cause it. There is a link to a separate video I recorded for adding your own f critical program to the calculator, it's in the description, you could try adding the program instead?

[@blaizakin90]

@@quantquill wow you responded fast thank you so much. I think I figured it out though, it has something to do with the fact I'm using the plus ce basically I need to click the graph button for it to calculate. Pretty much it just tells you some pre-inputted value for your variable if you don't click graph. No clue why it is the graph button that does it, but it is apparently.