Author Topic: What is a confidence interval?  (Read 460 times)

tweeter

  • Getbig IV
  • ****
  • Posts: 2180
What is a confidence interval?
« on: August 17, 2011, 05:37:18 PM »
Can someone give me a quick, summarized answer as to what confidence interval means. Everything I look up online goes into too much depth. Like when someone quotes a statistic within a given confidence interval, what does this mean?

burn2live

  • Getbig IV
  • ****
  • Posts: 3698
Re: What is a confidence interval?
« Reply #1 on: August 17, 2011, 06:02:00 PM »
Can someone give me a quick, summarized answer as to what confidence interval means. Everything I look up online goes into too much depth. Like when someone quotes a statistic within a given confidence interval, what does this mean?

Basically a range of values centred on a point estimate which are likely to include the population parameter, with a given probability. Usually the probability selected is 95%, therefore 95% of the confidence limits of random samples would capture the specified population parameter

So if you took 20 blood glucose measurements (the random samples) and calculated the CIs, in theory 19 of the sample confidence limits would overlap with the population mean (the specified population parameter) and 1 would not.

Hope this helps a bit. Statistics is rough  :(

The True Adonis

  • Getbig V
  • *****
  • Posts: 50255
  • Fear is proof of a degenerate mind.
Re: What is a confidence interval?
« Reply #2 on: August 17, 2011, 06:07:10 PM »
Confidence Interval- The limited amount of time you have to impress either females or males with a pumped up muscle.  The Confidence Interval of the Pump is unknown, yet it drops exponentially as time passes. (this can be remedied however by excusing yourself to do pushups in the bathroom or dip between stall doors)

Mr. Magoo

  • Getbig V
  • *****
  • Posts: 9808
  • THE most mistaken identity on getbig
Re: What is a confidence interval?
« Reply #3 on: August 17, 2011, 06:12:16 PM »
Basically a range of values centred on a point estimate which are likely to include the population parameter, with a given probability. Usually the probability selected is 95%, therefore 95% of the confidence limits of random samples would capture the specified population parameter

So if you took 20 blood glucose measurements (the random samples) and calculated the CIs, in theory 19 of the sample confidence limits would overlap with the population mean (the specified population parameter) and 1 would not.

Hope this helps a bit. Statistics is rough  :(

this^