- Thread starter bhardwajsaurabh
- Start date

Please look at the formula for the confidence interval ...

CI = average ± Z_1-α/2 * SD/sqrt(n)

The value after the ± is the MOE.

So you have all the data you can take out the sample size n.

http://www.statskingdom.com/doc_confidence_interval.html

Please look at the formula for the confidence interval ...

CI = average ± Z_1-α/2 * SD/sqrt(n)

The value after the ± is the MOE.

So you have all the data you can take out the sample size n.

http://www.statskingdom.com/doc_confidence_interval.html

You are really doing a mess with the numbers... you change the numbers all the time

What is average? :0.1 , 120000, 129,517 1,290,517 ? you should put the comma every 3 digits.

what is the STD: 18300 20,914?

With a confidence level of 0.95 => α=0.05 => 1-0.05/2=0.975 => Z_0.975=1.96

You probably mean that the MOE=5% not 5. that is mean a percentage of the average.

To get an accuracy of +-5 with a standard deviation of 18300 you need a huge sample size ... 5% is a different story ...

Also is the population finite?

You are really doing a mess with the numbers... you change the numbers all the time

What is average? :0.1 , 120000, 129,517 1,290,517 ? you should put the comma every 3 digits.

what is the STD: 18300 20,914?

With a confidence level of 0.95 => α=0.05 => 1-0.05/2=0.975 => Z_0.975=1.96

You probably mean that the MOE=5% not 5. that is mean a percentage of the average.

To get an accuracy of +-5 with a standard deviation of 18300 you need a huge sample size ... 5% is a different story ...

Also is the population finite?

STD:209,14

MOE:-5%

N=12

CL-1.96(95%)

Hi Saurabh,

I thought the question is what is the required sample size???

commonly used:

N - population sample size, for example, id there are 10,000,000 voters in the country and you sampled 5000: N=10,000,000 n=5000.

I thought the question is what is the required sample size???

commonly used:

N - population sample size, for example, id there are 10,000,000 voters in the country and you sampled 5000: N=10,000,000 n=5000.

Yes my question is the desired sample size ,when average population is 129,517

Slowly slowly I will understand your question ...

It is better to support all the data at the beginning ... (but shortly...)

You should probably use the expected proportion of the thing you want to check., not the STD.

What is the STD you mention? STD of what?

http://www.statskingdom.com/50_ci_sample_size.html

Slowly slowly I will understand your question ...

It is better to support all the data at the beginning ... (but shortly...)

You should probably use the expected proportion of the thing you want to check., not the STD.

What is the STD you mention? STD of what?

http://www.statskingdom.com/50_ci_sample_size.html

So what was the problem ...?

Let me explain you the entire scenario.

I am sharing dummy data below:

In a call center we receive approx 0.1 million calls per month.we want to conduct a survey to gauge the c-sat.

Now i want to know,how many customers should i touch in a month (desired sample) for the survey.

Mean on 12 months call volume is 0.12 million,while SD of call volume in past 12 months is 18,000.

If you don't know the expect proportion you should take the worse case: Proportion=0.5

population=129,517

confidence level=0.95

MOE=0.05

sample size=385

This says that if you get the result that 0.5 (50%) of the sample prefer the green color.

The confidence interval will be 0.5±0.05 or 50% ± 5%.