Data considerations for Analyze Binary Response for Definitive Screening Design

To ensure that your results are valid, consider the following guidelines when you collect data, perform the analysis, and interpret your results.

You must have created or defined a definitive screening design in Minitab

A designed experiment in Minitab must have at least 2 factors that are either continuous or categorical.

The response variable should be binary

A binary response has two outcomes, such as pass or fail. Response data should be organized into two worksheet columns using the event/trial format. For more information, see Enter your data for Analyze Binary Response for Definitive Screening Design.

If your response is continuous, then you should use Analyze Definitive Screening Design.

Ensure that the measurement system produces reliable response data

If the variability in your measurement system is too great, your experiment may lack the power to find important effects.

Each observation should be independent from all other observations
If your individual observations are dependent, your results might not be valid. Consider the following points to determine whether your observations are independent:
  • If an observation provides no information about the value of another observation, the observations are independent.
  • If an observation provides information about another observation, the observations are dependent.
The experimental runs should be randomized

Randomization reduces the chance that uncontrolled conditions will bias the results. Randomization also lets you estimate the inherent variation in materials and conditions so that you can make valid statistical inferences based on the data from your experiment.

Collect data using best practices
To ensure that your results are valid, consider the following guidelines:
  • Make certain that the data represent the population of interest.
  • Collect enough data to provide the necessary precision.
  • Record the data in the order it was collected.
The model should provide a good fit to the data

If the model does not fit the data, the results can be misleading. In the output, use the residual plots, the diagnostic statistics for unusual observations, and the model summary statistics to determine how well the model fits the data.

    By using this site you agree to the use of cookies for analytics and personalized content.  Read our policy