22.Sample Size

Confidence Interval,

The uncertainty associated with sampling is defined by something called the Confidence Interval, which is also called the margin of error in some applications. You’ve likely read or seen survey or poll results reported with a margin of error: The fans are 93 percent in favor of the new team colors, +/- 2 percent. The margin of error in that particularly survey was 2 percent, or 0.02.

What Information is Required for Choosing Sample Size?

The alpha level you set – remember, the default in Minitab is typically set at 0.05

The beta level can be set by the experimenter and a sample size calculated from that number. If the sample size is fixed, then the experimenter usually sets alpha and calculates the beta risk from the sample size.

Delta: The practical difference the experimenter wants to detect using the test.

The estimated population standard deviation.

Typpe of Data, Discrete or continuous?

Which hypothesis test are you using

Choosing the Right Method

1-Sample Variance Test Sample Size Calculation

1-Sample Proportion Test Sample Size Calculation

2-Sample Variance Test Sample Size Calculation

2-Sample Proportion Test Sample Size Calculation

2-Sample T Test Sample Size Calculation

Analysis of Variance (ANOVA) Sample Size Calculation

1-Sample Z Test Sample Size Calculation

Design of Experiment (DOE) Sample Size Calculation

1-Sample T Test Sample Size Calculation

  • Means testing
  • Comparing to a target value
  • You already have sample statistics about the population
  • Means testing
  • Comparing to a target
  • You do not have sample statistics about the population (standard deviation is not known)

Means testing

Comparing means from two sets of data

  • Variance testing
  • Comparing variance from two sets of data
  • Variance testing
  • Comparing variance of one data set to a target

Proportion testing (rate, x per y)

Comparing rate of one data set to a target

  • Proportion testing (rate, x per y)
  • Comparing rate from two sets of data
  • Means testing

More than 2 sets of data

Only one factor for x

More than 2 sets of data

More than one factor for x

Means testing