Intro to Statistics at IU Indy - One-Sample Tests Lecture 2
Overview
Andy Baker explains one-sample hypothesis tests by comparing a calculated z or t statistic with a critical value at alpha = 0.05. He distinguishes one-tailed from two-tailed tests, demonstrates how to choose t-table values using degrees of freedom, and works examples involving grocery bills, library visits, and a student-biking proportion.
Key takeaways
- At alpha = 0.05, a one-tailed Z test uses a critical value of +1.645 or -1.645 depending on direction, while a two-tailed Z test uses ±1.96.
- For a one-sample t test, degrees of freedom are n - 1; with 35 degrees of freedom, the alpha = 0.05 critical value is 1.690 for a right-tailed test and approximately ±2.030 for a two-tailed test.
- In the grocery-bill example, n = 36, x̄ = $49, μ₀ = $40, and s = $18 produce a standard error of $3 and z = 3.0, exceeding the right-tail cutoff of 1.645.
- In the library-visits example, x̄ = 2.3, μ₀ = 2, s = 1.8, and n = 36 produce t = 1.00; because this is below the 1.690 cutoff, the null is not rejected.
- For a one-sample proportion Z test, the standard error is calculated from the null proportion; with p₀ = 0.20 and n = 100, it is 0.04, and p̂ = 0.10 gives z = -2.5.
- Failing to reject a null hypothesis does not establish that it is true; it means the sample did not provide sufficient evidence to reject it at the selected significance level.
Chapters
- Andy Baker moves the project proposal deadline from Sunday to the following Sunday, November 8, to give students more time to develop workable ideas.
- Lab 7 is available to review, with a supplemental video planned; Baker also encourages students to submit late or locked Labs 1 through 6 by email.
- He notes missing quiz scores and says students who can still view a locked assignment should complete it and send it directly rather than having it reopened.
- Baker plans to move first-pass concept explanations online, where students can pause and rewind, while using Thursday class meetings for worked problems and quiz preparation.
- Quizzes will happen every other meeting rather than every class; the next meeting will focus on problems for Quiz 9, scheduled for November 12.
- A single 20-point bonus quiz will be offered online near the end of the semester.
- For alpha = 0.05, the confidence level is 95%; alpha = 0.01 corresponds to 99% confidence.
- A two-tailed Z test at alpha = 0.05 uses critical values of -1.96 and +1.96, allowing for differences in either direction.
- A one-tailed Z test uses a critical-value magnitude of 1.645; the sign is positive for a right-tail test and negative for a left-tail test.
- For a one-sample t test, degrees of freedom equal n - 1, so a sample of 36 has 35 degrees of freedom.
- With 35 degrees of freedom, a two-tailed test at alpha = 0.05 uses the 0.025 column and has critical values of approximately ±2.030.
- The t distribution accounts for sample size, especially small samples, so its critical values differ from the fixed Z cutoffs.
- A test predicting a decrease is left-tailed; a test predicting an increase is right-tailed, and the critical-value sign follows that direction.
- For n = 18, degrees of freedom are 17; a one-tailed test at alpha = 0.05 has a critical-value magnitude of 1.740.
- A two-tailed test is often preferred in research because it tests for differences in either direction rather than prespecifying an increase or decrease.
- The claim is that the mean weekly grocery bill is $40, while the researcher suspects it is higher, making the test right-tailed.
- The sample has n = 36, a mean of $49, and a standard deviation of $18; Baker uses a Z statistic for this worked example.
- At alpha = 0.05, the right-tail Z critical value is +1.645; statistics above it fall in the rejection region.
- The standard error is s / √n = 18 / √36 = 3 dollars.
- The test statistic is z = (49 - 40) / 3 = 3.0, which exceeds the +1.645 critical value.
- The result falls in the rejection region, so the $40 null claim is rejected in favor of evidence that the mean is higher.
- For the grocery example, the appropriate decision is to reject the null hypothesis, with the conclusion stated in the context of the $40 mean claim.
- A statistic that does not cross the critical value leads to failing to reject the null, not proving that the null claim is true.
- Baker previews a t-test example about weekly campus-library visits and a proportion test about students who bike to campus.
- The claim is a mean of 2 campus-library visits per week, and the suspected increase makes the test right-tailed.
- With n = 36, degrees of freedom are 35; the one-tailed alpha = 0.05 t-table critical value is +1.690.
- Using x̄ = 2.3, μ₀ = 2, and s = 1.8 gives a standard error of 0.3 and t = (2.3 - 2) / 0.3 = 1.00.
- Because 1.00 is below 1.690, the result fails to reject the null despite the sample mean being slightly above 2.
- The population claim is that 20% of IU Indianapolis students bike to campus; suspecting a difference in either direction calls for a two-tailed test.
- For a proportion test, Baker uses a Z statistic rather than a t statistic, and the hypotheses use the population proportion p rather than the sample proportion p̂.
- The sample contains 10 bikers out of 100 students, so p̂ = 0.10, compared with the claimed p₀ = 0.20.
- The standard error is built from the null claim: √[0.20 × (1 - 0.20) / 100] = 0.04.
- The test statistic is z = (0.10 - 0.20) / 0.04 = -2.5.
- The two-tailed critical values are -1.96 and +1.96, so -2.5 lies in the rejection region and the 20% claim is rejected.
- Baker recaps that the class worked through a one-sample Z test for a mean, a t test for a mean, and a Z test for a proportion.
- The remaining meal-plan example was not completed in class, and Baker says he will try to cover it in an online video.
- Students are invited to email questions and project ideas before the next Thursday meeting.
Transcript excerpt
0:02 Yeah, you're gonna you're going to talk about that. We're gonna talk about that. That's That's the deal. Yo. All right. We fire it up here. There we go. All right. So, we're zoomed in here. Zoom it out. Got some changes and updates to the schedule. First off, um the proposal pushed it back from this Sunday to the
0:52 following Sunday. So, give us just a little bit more time to get some ideas. But that doesn't mean, oh, it's more time to kick the cane down the road. It's actually more time for you to make sure that you get something done that we actually can work with. So, I kind of I'm getting, you know, just the the scent uh that some people are kind of maybe um pretty far off uh on the proposal from getting there. So, uh that's why I pushed that back just a few. Um, so I already getting a couple um uh um project ideas and uh they both uh the ones I've heard so far, they have my toes at tapping and so they're great sounding ones. So we can do it. Uh I definitely think we can do it. Um as far as some things I want to mention, Lab 7
1:41 is available to you. Um I will try to maybe make a video here, so don't go off and do it just yet. Uh, but it is available to you as far as uh to peruse. It's also, I think, due November 8th as well. Um, so it's due November 8th as well. I could be wrong on that, but uh uh that one will have a probably a video I want to make tomorrow. We'll see how what happens as far as the weather, all that. Um, but that's going to be made
This is the start of the transcript. Save this video free for the whole transcript, searchable and time-coded.
Related notes
Summary, takeaways, and chapters were generated by AI from the video's transcript and may contain errors. The video belongs to its creator, Andy Baker.