2008 First Semester STATS课程官方问答贴

z-score

原帖由 punklink 于 2008-5-27 21:55 发表


which of the two tests from B) is more appropriate for these data?

paired, but how to justufy?

Question 5 dont know how to justufy them... ??

Question 4: sample of 20 differences are not symmetric, small sample size, can't use CLT, need to rely on the sign test, so it's the non-parametric test on the median.

Question 5: F-test not valid cause of the equal sd assumption not satisfied.

z-score

原帖由 ~Kurosagi~ 于 2008-5-27 21:43 发表
是
B
E
嗎??

不是E
B
嗎??

啊？不是B, E吗？第一个scenario的hypothesis是H0: p = 0.73.

璐秋孟孟

谢谢版主了！不过好像不用写太多

璐秋孟孟

还想请问4(c)应该是t-test还是sign test更合适呢？t-test好像只有slight skewness,除了有个outlier外没有其他问题了，不知道选哪个好

z-score

Question 7e
H0: mu(kfc)=mu(wendy's)=mu(roy rogers)=mu(burger king)
H1: There is at least one mean price of full meal that is different from the rest.
We have extremely strong evidence that one restaurant has a mean full meal price that is different from the rest.
f
Yes, interpret the P-value and CI.
Wendy's and the other three restaurants, list them out, don't write the answer like how I wrote it here.
Yes, Wendy's.

z-score

原帖由 璐秋孟孟 于 2008-5-27 22:09 发表
还想请问4(c)应该是t-test还是sign test更合适呢？t-test好像只有slight skewness,除了有个outlier外没有其他问题了，不知道选哪个好

我用explore在SPSS算了一下skewness statistic, 好像是-1.something, 意味着skewness不小。

z-score

原帖由 璐秋孟孟 于 2008-5-27 22:06 发表
谢谢版主了！不过好像不用写太多

是的

punklink

原帖由 z-score 于 2008-5-27 22:01 发表


Question 4: sample of 20 differences are not symmetric, small sample size, can't use CLT, need to rely on the sign test, so it's the non-parametric test on the median.

Question 5: F-test n ...

i see.. thanks ATM!!

wongchuntat905

thx~~~ of u~~ i hv just finished my assignment~ thx for all those effort u hv done to us~~~

Cheers~~~

U ar the best~

punklink

原帖由 z-score 于 2008-5-27 22:12 发表


是的

justify ... thank you so mcuh!

拂洛依德

你给的答案是题目改过之后还是没改？这是题目改过的－－〉Ｑ７（ｉｉ）Between which pair (or pairs) of chains were there significant differences
(at the 5% level) in the mean difference in price for a full meal?" 我们还有问Ｑ７（ｃ）．．．谢谢麻烦了！

punklink

7c?
still cant get ratio of standard deviation of most variable and least variable

ldp

108 Question 5 (c)

是不是 NO 啊！ 因为有outlier，所有用median比较好？ 对不对？

璐秋孟孟

谢谢啊！那skewness是moderate 还是算strong呢？

luvwei

7 d please

莣卻啲J朢

同样问5c 有outlier是不是还有cluster啊
那不用t test 用什么代替呢?

~Kurosagi~

原帖由 punklink 于 2008-5-27 22:19 发表
7c?
still cant get ratio of standard deviation of most variable and least variable

most variable is the largest range which is Wendy's

the least vairable is the shortest range which is BK

z-score

原帖由 拂洛依德 于 2008-5-27 22:18 发表
你给的答案是题目改过之后还是没改？这是题目改过的－－〉Ｑ７（ｉｉ）Between which pair (or pairs) of chains were there significant differences
(at the 5% level) in the mean difference in price fo ...

啊？题目改动了？我都不知道，我没有enrol这门课所有没有收到通知。

z-score

原帖由 punklink 于 2008-5-27 22:19 发表
7c?
still cant get ratio of standard deviation of most variable and least variable

Can't quite remember the numbers, it's something like 0.37 divide by 0.1, the sd for wendy's and the sd for burger king.

charlie2e

谢谢阿童木大大真的帮了好大的忙！

z-score

原帖由 莣卻啲J朢 于 2008-5-27 23:00 发表
同样问5c 有outlier是不是还有cluster啊
那不用t test 用什么代替呢?

Question 5 part c问t-test是不是appropriate, 我们不用写怎样代替，说他不appropriate就好了。

z-score

原帖由 charlie2e 于 2008-5-27 23:12 发表
谢谢阿童木大大真的帮了好大的忙！

谢谢大家的支持，不客气，不客气。

z-score

原帖由 ldp 于 2008-5-27 22:22 发表
108 Question 5 (c)

是不是 NO 啊！ 因为有outlier，所有用median比较好？ 对不对？

恩，是的，outliers加上not normal. T-test on the difference of two means is not appropriate for the data.

z-score

原帖由 拂洛依德 于 2008-5-27 22:18 发表
你给的答案是题目改过之后还是没改？这是题目改过的－－〉Ｑ７（ｉｉ）Between which pair (or pairs) of chains were there significant differences
(at the 5% level) in the mean difference in price fo ...

哦，咳，原来是拼写错误，没事，我都没看出来，做法还是一样的。

z-score

好，阿童木准备睡了，好困阿。

~Kurosagi~

辛苦了謝謝!!!!!!

punklink

原帖由 z-score 于 2008-5-27 23:20 发表
好，阿童木准备睡了，好困阿。

thanks, good night!

spirithell

S1 2007 Exam
Q1. To display the percentage of claims for each age group, the most appropriate
tool to use would be:
(1) a scatter plot.
(2) a histogram.
(3) a bar graph.
(4) back-to-back stem-and-leaf plots.
(5) side-by-side box plots.
The Answer is (3), I wonder why not (2)?  Because for Q2 see below

2. Which one of the following statements is false?
(1) We cannot tell if there are any outliers for the cost of claims from Table 1
alone.
(2) Age Group is an ordinal variable.
(3) Approximately 69.5% of all claimants were aged 21–50.
(4) The interquartile range for claims made by those aged 61–70 is $81.80.
(5) A histogram would be an appropriate plot to display the cost of all claims.

The answer is (4), So that means the (5) is correct! And It also can be right in Q1.....

Help PLS.

Q5. Under the null hypothesis, the expected count for the 21–30 age group is
approximately 3123 and this age group’s cell contribution to the Chi-square
test statistic is approximately:
(1) 708
(2) 0.48
(3) 1487
(4) 0.32
(5) 480


Conditions:
Tatol Claim: 18,590  
population by Age Group(21--30): 16.8%
Claimants by Age Group(21-30): 24.8%

The Answer is (1)708! I think its wrong !! Why ？I think sjould be (3)1487.


PLS HELP!!!! URGENT!
by the waY WHAT IS YOUR GENDAR?

z-score

原帖由 spirithell 于 2008-5-28 16:42 发表
S1 2007 Exam
Q1. To display the percentage of claims for each age group, the most appropriate
tool to use would be:
(1) a scatter plot.
(2) a histogram.
(3) a bar graph.
(4) back-to-back st ...

I see, good question. Good effort for starting your revision so early.

Question 1 we have to use a bar graph because the percentage for age groups involves displaying data for a qualitative variable. For qualitative variables, we use bargraphs.

Question 2, the cost is a quantitative variable, therefore we can use a histogram.

z-score

原帖由 spirithell 于 2008-5-28 16:42 发表
S1 2007 Exam
Q1. To display the percentage of claims for each age group, the most appropriate
tool to use would be:
(1) a scatter plot.
(2) a histogram.
(3) a bar graph.
(4) back-to-back st ...

cell contribution = (observed - expected)^2/expected

Try this formula.

I'm ATM, hence my gender is.