topic Re: Help me verify answers by using SAS output in Statistical Procedures

Help me verify answers by using SAS output

JaeGonKim — Fri, 13 Oct 2017 08:52:22 GMT

Hi Everyone,

I am learning SAS so I can not solve this example. I want to test it on SAS to help my students understand. Can you help me solve by using SAS the example below?

Thanks,
Jae

Re: Help me verify answers by using SAS output

Ksharp — Tue, 10 Oct 2017 11:46:00 GMT

You can use nonparameter statistical method like resampling .

Check @Rick_SAS blog .

https://blogs.sas.com/content/iml/2014/11/21/resampling-in-sas.html

Re: Help me verify answers by using SAS output

Rick_SAS — Tue, 10 Oct 2017 11:56:10 GMT

You can use PROC TTEST to perform a one-sample test for the mean. You can use the SIDES=U option to perform a one-sided test and the H0= option to specify the value of the null hypothesis. There is a Getting Started example in the PROC TTEST documentation that shows how to perform the test and analyze the results.

data Movie;
input rating @@;
datalines;
3 5 4 3 4 5 5 4 3 2 5 4 3 2 3 5
;

proc ttest data=Movie H0=4 sides=u alpha=0.05;
var rating;
run;

Re: Help me verify answers by using SAS output

JaeGonKim — Tue, 10 Oct 2017 11:56:50 GMT

Hi Ksharp,

It is very difficult for me, please try to help me. If you can make SAS file and post it on here, I am really grateful to you.

Joe

Re: Help me verify answers by using SAS output

Ksharp — Tue, 10 Oct 2017 12:00:03 GMT

@Rick_SAS,

and I quoted "without using statistical software." .

Re: Help me verify answers by using SAS output

JaeGonKim — Tue, 10 Oct 2017 12:04:38 GMT

Hi Rich_SAS,

Thank you for your help.

Because SAS is quite new to me, could you send me a SAS file for this example and this results to email? My email is: jaegonkim.edu@gmail.com

Joe

Re: Help me verify answers by using SAS output

Rick_SAS — Tue, 10 Oct 2017 12:05:09 GMT

The problem says "(a) through (c) without using software," which I assume the OP can do. The OP was asking about part (d), "Verify your answers by using SAS output."

Re: Help me verify answers by using SAS output

Rick_SAS — Tue, 10 Oct 2017 12:06:20 GMT

The SAS code is posted in my response. You can copy and paste from this discussion thread.

Re: Help me verify answers by using SAS output

JaeGonKim — Tue, 10 Oct 2017 12:11:09 GMT

Hi Rick_SAS,

Yes, thank you. Could you solve (a) to (c) without using software? I want to know the results are to make a comparison of them.

Re: Help me verify answers by using SAS output

Ksharp — Tue, 10 Oct 2017 12:14:18 GMT

Sure. Rick .

Re: Help me verify answers by using SAS output

Rick_SAS — Tue, 10 Oct 2017 12:19:17 GMT

?? When you asked this question, you said you wanted to use SAS "to help my students understand," so I assumed you are a statistics teacher. This problem is solved in every elementary statistics textbook. I suggest you refer to the one that you are using for your class.Look in the chapter that discusses "tests of hypotheses" and the section that discusses "tests for the mean."

Re: Help me verify answers by using SAS output

JaeGonKim — Tue, 10 Oct 2017 12:29:44 GMT

Hi Rick_SAS,

Sorry, I will review it. I would like to ask only another example about SAS. Are you ready to help me?

Thank you.

Joe

Re: Help me verify answers by using SAS output

JaeGonKim — Tue, 10 Oct 2017 12:30:03 GMT

Hi Rick_SAS,

Sorry, I will review it. I would like to ask only another example about SAS. Are you ready to help me?

Thank you.
Joe

Re: Help me verify answers by using SAS output

Rick_SAS — Tue, 10 Oct 2017 12:33:52 GMT

If it is a different question, please start a new thread.

There are many people on the forum who can help answer your questions about how to use SAS to analyze data and how to interpret SAS output. I am just one of hundreds of people who answer questions on this community.

Re: Help me verify answers by using SAS output

Ksharp — Tue, 10 Oct 2017 12:51:32 GMT

I would like to give it a try.

a) H0 : mean=4 H1: mean>4

b) the statistic is the number of obs>4 , which is 5

c) due to 5/nobs = 0.312 , i.e. 0.3 > 0.05 which means can't refuse H0 .

d) from the output of proc ttest, can verify the above. You can't say mean at least 4.

SAS Output

均值	95% 置信限均值		标准差	95% 置信限标准差
3.7500	3.2834	正无穷	1.0646	0.7864	1.6476

自由度	t 值	Pr > t
15	-0.94	0.8188

Re: Help me verify answers by using SAS output

JaeGonKim — Tue, 10 Oct 2017 13:02:01 GMT

Hi Ksharp,

b. the value of the test-statistic is computed = 5. Right?, could you explain this value?

Re: Help me verify answers by using SAS output

Rick_SAS — Tue, 10 Oct 2017 13:11:44 GMT

With respect, I do not think that KSharp's analysis is correct. If you want to perform a one-sided test of the hypothesis that mu>4 against the alternative that mu<=4, then read the excellent online statistics course from PennState.

In particular, the test statistic should be t = (x_bar - mu0)/StdErr.

Unfortunately, I have other things to do, but I wish you luck in your efforts.

Ho and Ha

JaeGonKim — Sun, 15 Oct 2017 16:50:03 GMT

Hi everyone,

I have a example:

A movie was rated by the viewers who had watched it. The scores range from a low of 1 (very uninteresting) to a high of 5 (very fascinating). A sample of 16 viewers was chosen and their ratings are listed below:

3 5 4 3 4 5 5 4 3 2 5 4 3 2 3 5

Suppose the ratings of the movie can be assumed to be normally distributed. The movie production company claimed that the mean rating score is at least 4. An independent movie review company wants to test this claim. Answer questions (a) through (c) without using statistical software. Using α=0.05, would the movie review company conclude that the mean rating score of the movie is at least 4?.

Could you help me tell what is the null and alternative hypotheses: Ho vs Ha?

Thanks,

Joe

Re: Ho and Ha

Reeza — Sun, 15 Oct 2017 17:01:44 GMT

You may find this helpful:

https://support.sas.com/edu/schedules.html?crs=STAT1&ctry=us

Note that the e-course is free.

Re: Ho and Ha

JaeGonKim — Sun, 15 Oct 2017 17:06:24 GMT

Hi Reeza,

I know, but I don't sure. Ho >= 4 or > 4?. Please tell me know.

Thanks,

Joe