Enjoy learning Statistics Online! Please be sure to share and subscribe to our YouTube channel.

ASK THE PROFESSOR FORUM

Course Documents

Chapter 1 - Intro

Chapter 2 - Methods for Describing Sets of Data

Chapter 3 - Probability

Chapter 4 - Discrete Random Variables

Chapter 5 - Normal Random Variables

Chapter 6 - Sampling Distributions

Chapter 7 - Confidence Intervals

Chapter 8 - Tests of Hypothesis: One Sample

Chapter 9 - Confidence Intervals and Hypothesis Tests: Two Samples

Sample Exam I: Chapters 1 & 2

Sample Exam II: Chapters 3 & 4

Sample Exam III: Chapters 5 & 6

Sample Exam IV: Chapters 7 & 8

Ask the Professor Forum

Hi Professor

I can't seem to get questions #18 on chapter 14 homework. I was wondering if you could help me with the problem, I think I'm doing the ranking portion wrong.

Thanks

Posted to STATS 2 on Tuesday, December 2, 2014   Replies: 4


Professor Mcguckian
12/02/2014
5:35 PM EST

Hi Sandra,

We do not rank the data for the sign test, so if you ranked the data, you are definitely going wrong in that place. All you need to do is to count the number of values below zero and the number above zero, because the claim says the median is equal to zero. S- = the amount of values below zero = 4, and S+ = the amount of values above zero = 11. For a two-tailed test like this one, the test statistic is the larger of these two values (i.e., Max(4,11) = 11). This makes our test statistic equal to 11. 

I hope that answers your question,

Professor McGuckian 


.
12/02/2014
10:13 PM EST
Well now that makes a lot of sense!
Thank you!

.
12/03/2014
10:44 AM EST
Im doing some practice problems and I just came across question #36 on Ch. 14 homework, I don't know what the claim would be

Use the data below, the Kruskal-Wallis H-test, and a 2.5% significance level to construct an
ANOVA table to test the claim that there is a significant difference between the bonding
strengths.

Does that mean the claim is not equal to, or they're all equal?

Thank you again!

Professor Mcguckian
12/03/2014
11:08 AM EST

Hi Sandra,

The claim for the test here is basically the same as the alternative hypothesis. Remember that the Kruskal-Wallis H-test always uses the same null and alternative hypothesis pair:

Ho: The medians are all the same.

Ha: At least two of the medians differ from each other significantly

Post a reply / comment below or click here to submit a new question to the forum.