Friday, November 21

Message: 25032
Posted by: Eric G
Posted on: Thursday, 20th March 2003

---

Can someone please explain why you would choose one of these tests over the other.  2 proportion, 2 sample T test .

Why compare for mean or variance?

 Laymans terms please would definitely would help.

Thanks

---

Message: 25052
Posted by: Sam
Posted on: Friday, 21st March 2003

---

Your question is - Why compare for mean or variance?

                Several reasons:

Did your improvement make a difference? Campare your process mean before & after change.

If your are looking at Subgroup to Subgroup to see the difference.

(Example : Which carrier has the best delivery time? You can compare the mean of each carrier and the amount of variation.)

my 2 cents.

---

Message: 25054
Posted by: Jamie
Posted on: Friday, 21st March 2003

---

First ask yourself what you want to know... then determine the test you should use to answer that question. You need to form the question as a hypothesis. 

The two tests you are asking about apply to testing differences between 2 groups. So, do you want to know if the mean of the two groups is different or is the variance of the two groups different. If its the means of continuous normally distributed data you should use the 2 sample t-test. You might first test if these 2 groups have different variances because you can you the pooled std dev for the groups in the 2 sample t-test to make the test more powerfull.

If your data is in the form of percents..., for example if I want to test the percent of defectives for shift 1 vs shift 2, then I would you 2 proportions test.

It all depends on what you want to know and what form your data is in.

Jamie

---

Message: 25057
Posted by: Chris
Posted on: Friday, 21st March 2003

---

The tests are different in the respect that the tests for proportions are used for discrete data and t-tests are for continuous data.

For example: You would use the test of proportions to determine if the fraction nonconforming between 2 different processes are equal. You would use the 2 sample t-test to test for a difference in two populations means (diameters, lengths, etc.)

Neither test would be used to compare variances. To test variances you would use the F test, Chi-square, Barlett's test, and others.

---

Message: 25060
Posted by: Jamal Yamak
Posted on: Friday, 21st March 2003

---

Chris,

Great answer.

Jamal

---

Message: 25061
Posted by: Steve G
Posted on: Friday, 21st March 2003

---

Eric,

Chris hit the nail on the head re. your specific question, but before I looked for root cause in my variance and mean I would ask two other questions:

1. Is the process stable over time?
2. What is the shape of the data (normal/non-normal)? 

The reason for these questions is that checking process stability (by run chart or control chart) may immediately throw up special cause and that the shape/normality of the data will affect your choice of analysis tool.  After this you can test for root cause by seeing if the stratification criteria within the factors that you are analysing are differently spread or are centred at different points. 

I remember this analysis process as STABILITY, SHAPE, SPREAD, CENTRE.  As some tests of centring assume equal variance then you should always analyse spread before centring. 

This may be more than you were after but I'll give you the specific tools for Minitab analysis.  

• If you are looking at discrete data you can use proportions tests or Chi-square. 
• If you are looking at normally distributed data then you can use 2-Variances test for variance and t-tests for means. You can use ANOVA; this has the advantage of not only showing you if a factor is statistically significant but shows the proportion of the total variance that is due to your factor.
• If you are looking at non-normal data then use Test of Equal Variance (a.k.a. Homogeneity of Variance) and Mood's Median Test.  You could use the GLM (Generalised Linear Model) in place of ANOVA.   

Good hunting! 

Steve

---

Message: 25076
Posted by: ERIC
Posted on: Friday, 21st March 2003

---

Thanks to all of you. One of these days I will be knowledgeable enough to provide insight to others. I think is called "paying it forward"

---

Message: 48404
Posted by: Leo Chan
Posted on: Tuesday, 22nd June 2004

---

Steve,

I have try to use a mood's median test to two sample, one normal and the other non-normal, then for the same samples, I have used a t-test, and I have 2 results different. Can you tell me which test to use in this case?

Thanks

---

Message: 70565
Posted by: A
Posted on: Friday, 20th May 2005

---

I wat the diff bw 2 prop and chi sq... not 2 pro and 2 sample...

---

Message: 70567
Posted by: Preeti Malik
Posted on: Friday, 20th May 2005

---

Ok here is the dofference :

Use the 2 Proportions  to compute a confidence interval and perform a hypothesis test of the difference between two proportions. For example, suppose you wanted to know whether the proportion of consumers who return a survey could be increased by providing an incentive such as a product sample. You might include the product sample with half of your mailings and see if you have more responses from the group that received the sample than from those who did not. For a two-tailed test of two proportions:

H0: p1 - p2 = p0 versus H1: p1 - p2 �‚ p0

where p1 and p2 are the proportions of success in populations 1 and 2, respectively, and p0 is the hypothesized difference between the two proportions

 

Chi Square :The Chi Square Test is a statistical test which consists of three different types of analysis 1) Goodness of fit, 2) Test for Homogeneity, 3) Test of Independence.

The Test for Goodness of fit determines if the sample under analysis was drawn from a population that follows some specified distribution.

The Test for Homogeneity answers the proposition that several populations are homogeneous with respect to some characteristic.

The Test for independence (one of the most frequent uses of Chi Square) is for testing the null hypothesis that two criteria of classification, when applied to a population of subjects are independent. If they are not independent then there is an association between them.

Chi Square is the most popular discrete data hypothesis testing method

 

Hope this helps

 

---

Message: 104324
Posted by: Sarika
Posted on: Saturday, 28th October 2006

---

Test of proportion is used to calculate the difference in two groups of "BINOMIAL" data in whereas T- test is performed on "VARIABLA" data set to measure (if any) the difference between two data sets.

---

Message: 123832
Posted by: Dilshad Botani
Posted on: Sunday, 12th August 2007

---

Dear Iric,

I can answer your question by two examples:

1. Unemployment rates in 2 Counties.  Is there a difference in the unemployment rates?               Test at a = .05.

County A:  100/400 Unemployed   (25%)

County B:  44/200 Unemployed    (22%)

2. if we have to the records of two samples students.

x1= 90, 59, 78,...,89

x2= 50, 70, 76,...,45

Is there a difference in the records of these two groups? 

Of course, we obliged to use proportion test with the first example. and t test for the second.

with regards

---