# Select the options for Stability Study

Stat > Regression > Stability Study > Stability Study > Options

The available options depend on whether you have a batch factor, and whether the batch factor is a fixed factor or a random factor.

## No batch factor

Percent of response within spec limits with specified confidence
• 50 percent (corresponds to mean response): Calculate the shelf life based on the mean of the response values. Minitab calculates the percentiles based on the normal distribution, so the 50th percentile is equal to the mean.
• Other percent: Calculate the shelf life based on the percentile that you specify.
Confidence level
Enter the level of confidence for the confidence interval.
Use one spec limit to calculate shelf life
Select this option when the response variable has two specification limits, but only one limit is relevant for the shelf life. The selection adjusts the confidence level to produce a one-sided bound when you enter two specification limits. For example, if the confidence level is 95% and you select this option, then the confidence level for the single bound that is relevant to the shelf life is also 95%. If you do not select this option, then the confidence level for both bounds is 97.5%. The single 95% bound gives a longer shelf life.
For example, a medication has an upper specification limit of 12.5 micrograms and a lower specification limit of 12 micrograms. In storage, the medication degrades, but never increases in strength. Only the lower specification limit is relevant to the shelf life.

## Batch is a fixed factor

Batch is a fixed factor
If you did not randomly select the batches for your study, the batch is a fixed factor. For more information, see Fixed versus random batches in Stability Study.
Percent of response within spec limits with specified confidence
• 50 percent (corresponds to mean response): Calculate the shelf life based on the mean of the response values. Minitab calculates the percentiles based on the normal distribution, so the 50th percentile is equal to the mean.
• Other percent: Calculate the shelf life based on the percentile that you specify.
Confidence level
Enter the level of confidence for the confidence interval.
Alpha for pooling batches
Enter the level of significance for model selection. During model selection, Minitab evaluates first the batch by time interaction, and then the batch term. If the p-value for the term is greater than or equal to the alpha-level you specify, then the term is removed from the model. Otherwise, the term remains in the model and model selection ends.
Use one spec limit to calculate shelf life
Select this option when the response variable has two specification limits, but only one limit is relevant for the shelf life. The selection adjusts the confidence level to produce a one-sided bound when you enter two specification limits. For example, if the confidence level is 95% and you select this option, then the confidence level for the single bound that is relevant to the shelf life is also 95%. If you do not select this option, then the confidence level for both bounds is 97.5%. The single 95% bound gives a longer shelf life.
For example, a medication has an upper specification limit of 12.5 micrograms and a lower specification limit of 12 micrograms. In storage, the medication degrades, but never increases in strength. Only the lower specification limit is relevant to the shelf life.

## Batch is a random factor

Batch is a random factor
If you randomly select the batches for the study, the batch is a random factor. For more information, see Fixed versus random batches in Stability Study.
Percent of response within spec limits with specified confidence
Percent
Calculate the shelf life based on the percentile that you specify. When the batch is a random factor, by default, Minitab use the 95th percentile, instead of the 50th percentile.
Confidence level
Enter the level of confidence for the confidence interval.
Alpha for pooling batches
Enter the level of significance for model selection. During model selection, Minitab evaluates first the batch by time interaction, and then the batch term. If the p-value for the term is greater than or equal to alpha-level you specify, then the term is removed from the model. Otherwise, the term remains in the model and model selection ends.
Use one spec limit to calculate shelf life
Select this option when the response variable has two specification limits, but only one limit is relevant for the shelf life. The selection adjusts the confidence level to produce a one-sided bound when you enter two specification limits. For example, if the confidence level is 95% and you select this option, then the confidence level for the single bound that is relevant to the shelf life is also 95%. If you do not select this option, then the confidence level for both bounds is 97.5%. The single 95% bound gives a longer shelf life.
For example, a medication has an upper specification limit of 12.5 micrograms and a lower specification limit of 12 micrograms. In storage, the medication degrades, but never increases in strength. Only the lower specification limit is relevant to the shelf life.
Provide starting estimates for variance components
If you don't specify starting estimates, Minitab uses the minimum normal quadratic unbiased estimation (MINQUE) estimates. The MINQUE estimates are usually good estimates.
Batch
Enter a starting estimate for the variance component of the batch factor.
Batch x Time
Enter a starting estimate for the variance component of the batch by time interaction (Batch x Time).
Error
Enter a starting estimate for the variance component of the error variance.
Maximum number of iterations
Enter the maximum number of iterations for estimating the unknown parameters in the models.
By using this site you agree to the use of cookies for analytics and personalized content.  Read our policy