so my question is am I supposed to purchase the toolpak in order to compute my answers because I am have a hard time here gaining my answers> Signed confused Giorvanna
Anbu: You're welcome. The 293 figure came from a separate, earlier study that I looked up at the time I made the video. I don't have a citation, and it may no longer be an accurate figure. The main thing to understand is that you use a one-sample t-test when you have such a know figure and want to compare a sample's average to it. Imagining a different scenario might help. Suppose, for example, that you're a quality control engineer in a factory that produces canned soup. At one point in the manufacturing process, the cans are filled with water that is supposed to be at a temperature of at least 100 degrees Celsius (212 degrees Farenheit). The "100 degree" figure is called for by the factory's standards, the recipe, government regulations, or some other external source. After selecting a random sample of water-filled cans from that point of the manufacturing process, you could measure the temperature of each, find the average, and compare it to 100 degrees Celsius using a t-test.
Heather: Technically, I *should* have chosen the one-tailed p-value, given that the hypothesis was overtly directional. In other words, I was hypothesizing that Americans' TV viewing was significantly *greater* than Swedes' TV viewing. That's exactly what the one-tailed p-value is for. I went with the two-tailed p-value, though, because, in many cases, one has no good reason for theorizing about the direction of the difference (that is, about whether the test average will be greater than the sample average, or about whether the test average will be less than the sample average). All one suspects is that the test average will be significantly *different* from the sample average. In this sense, the two-tailed p-value is more conservative and more broadly applicable. But your are quite right; the one-tailed p-value would be the correct choice in this particular scenario. Good comment.
I loved your trick! It saved me quite some time! Thanks.
Great explanation of the t-test!!
Thanks for the explanation random stranger. very helpful.
This was very helpful. Thanks very much.
Great video! Thanks!
Well said!
Hi Ken!! Can you tell me what is the null hypotheses and the alternate hypotheses here
Thanks
where does 293 come from ?thanks
so my question is am I supposed to purchase the toolpak in order to compute my answers because I am have a hard time here gaining my answers> Signed confused Giorvanna
first of all thanks to you.... but i have little confusion about where does 293 come from ?
Anbu: You're welcome. The 293 figure came from a separate, earlier study that I looked up at the time I made the video. I don't have a citation, and it may no longer be an accurate figure. The main thing to understand is that you use a one-sample t-test when you have such a know figure and want to compare a sample's average to it. Imagining a different scenario might help. Suppose, for example, that you're a quality control engineer in a factory that produces canned soup. At one point in the manufacturing process, the cans are filled with water that is supposed to be at a temperature of at least 100 degrees Celsius (212 degrees Farenheit). The "100 degree" figure is called for by the factory's standards, the recipe, government regulations, or some other external source. After selecting a random sample of water-filled cans from that point of the manufacturing process, you could measure the temperature of each, find the average, and compare it to 100 degrees Celsius using a t-test.
Why did you choose the p-value from the two-tailed test instead of the one tailed test in the output?
Heather: Technically, I *should* have chosen the one-tailed p-value, given that the hypothesis was overtly directional. In other words, I was hypothesizing that Americans' TV viewing was significantly *greater* than Swedes' TV viewing. That's exactly what the one-tailed p-value is for. I went with the two-tailed p-value, though, because, in many cases, one has no good reason for theorizing about the direction of the difference (that is, about whether the test average will be greater than the sample average, or about whether the test average will be less than the sample average). All one suspects is that the test average will be significantly *different* from the sample average. In this sense, the two-tailed p-value is more conservative and more broadly applicable. But your are quite right; the one-tailed p-value would be the correct choice in this particular scenario. Good comment.
Ken Blake Thank you!