Statistics help please

[mwayne]

Long story short, my employer has a tuition reimbursement program for approved classes so I'm taking a manufacturing process class. One of my homework problems has me confused and I would appreciate some help steering me in the right direction. The problem is from Niebel's Methods, Standards, and Work Design 13th Edition Chapter 14 problem 7 and is:
"If the average handling activity during a 10-day study is 82 percent, and the number of daily observations is 48, what is the confidence interval for each day's percentage activity?"
So I blew the dust off my statistics textbook and tried to figure it out myself but the best I can come up with is to assume a confidence interval and then calculate the margin of error. I must be missing something but have run out of ideas. To provide a bit of context for the chapter I'm working on, the previous question was to determine the minimum number of random samples to provide a 95% confidence interval. Thanks in advance!! This is a writing intensive class so I will cite you as a source if you can help.

 

[AWSubie]

Go online and get a Chegg account. My wife just finished a statistics course and Chegg is fantastic. Also Minitab software is very helpful.

 

[drssg]

It's been quite a while since my statistics classes, but it seems to me that the problem definition is incomplete, or perhaps there are other assumptions and/or data that are intended to be applied to the problem.

I believe a confidence interval has to have a percentage associated with it to have any meaning.

To calculate a confidence interval, I would think you need the raw data or be provided with the standard deviation of the data. The greater the standard deviation, the wider the confidence interval.

 

[mwayne]

Go online and get a Chegg account. My wife just finished a statistics course and Chegg is fantastic. Also Minitab software is very helpful.

Thanks AW, I was already aware of Chegg and back when I took statistics looked into a membership. At the time it seemed like one of those deals like with satellite TV where it might be difficult to stop the autobilling to the credit card so I passed on it. This class only has a couple more statistics questions so I'll pass on it again. How does billing work for your wife? Does she pay a monthly fee or a one time payment good for the rest of the semester?

 

[mwayne]

It's been quite a while since my statistics classes, but it seems to me that the problem definition is incomplete, or perhaps there are other assumptions and/or data that are intended to be applied to the problem.

I believe a confidence interval has to have a percentage associated with it to have any meaning.

To calculate a confidence interval, I would think you need the raw data or be provided with the standard deviation of the data. The greater the standard deviation, the wider the confidence interval.

That is my thinking also, confidence level is specified by management or the customer, not determined by field results. I'm going to start my answer by stating: "Assuming a 95% confidence interval......". And see what happens.
Thanks.

 

[Torvy]

It's been quite a while since my statistics classes, but it seems to me that the problem definition is incomplete, or perhaps there are other assumptions and/or data that are intended to be applied to the problem.

I believe a confidence interval has to have a percentage associated with it to have any meaning.

To calculate a confidence interval, I would think you need the raw data or be provided with the standard deviation of the data. The greater the standard deviation, the wider the confidence

I was thinking along the same lines. I am better at statistical analysis than the underlying math...it seems, however that you need to know 82% of what? If producing a billion widgets a day, 48 is a miniscule sample. If you produce 60 widgets a day 48 is huge. Is this a follow up question?

 

[AWSubie]

How does billing work for your wife? Does she pay a monthly fee or a one time payment good for the rest of the semester?

She paid for two months at $15/month.

We had no problem ending the subscription.

 

[mwayne]

I was thinking along the same lines. I am better at statistical analysis than the underlying math...it seems, however that you need to know 82% of what? If producing a billion widgets a day, 48 is a miniscule sample. If you produce 60 widgets a day 48 is huge. Is this a follow up question?

This is a work sampling question, getting a sample every 10 minutes throughout a 8 hour (480 minute) day and extrapolate results from that. I quoted the question verbatim from the book and it doesn't say what the 82% is. And it is not a follow up question, it may be a character building question lol!

 

[mwayne]

She paid for two months at $15/month.

We had no problem ending the subscription.

Thanks, I'll keep that in mind for the future.

 

[Torvy]

This is a work sampling question, gettisampling ng a sample every 10 minutes throughout a 8 hour (480 minute) day and extrapolate results from that. I quoted the question verbatim from the book and it doesn't say what the 82% is. And it is not a follow up question, it may be a character building question lol!

I have lived that life in a high production environment. (Producing bank statements and credit cards). We would not be able to calculate confidence level without knowing the throughput or the sample as a % of total volume. The way I read the 82% would be 'uptime' or the percent of the working hours that production is being run. In the case of statements you could run 30k in an 8 hour shift. Cards, more like 2400. Your sample size is based on time, so it seems you need some idea of the throughput. Or, you put together a formula with X being the missing volume.