How Does the Sample Set Process Work?

When creating a Sample Set, the user first chooses the type of sample to generate (Basic, Focused, or Random).

For each choice, the program displays and defaults to the suggested number of applications to pick based on USDA regulations.

The system prompts for a 'Random Seed' value, which is defaulted to a randomly generated number. If a specific random seed value is entered, then a repeatable sample selection is possible, whereby the same sample can be generated in a test database before generating it in a live database (for Advanced Users or troubleshooting only).

 1. The actual selection process works by first analyzing every student to determine whether or not they are pickable for verification. The determination is based on current eligibility and school type.

A student will not be picked if:

• Student is inactive

• Student is enrolled in a school that has been configured as a non-pickable type (Provision 2, non-participating, institutionalized)

• Student's current (or future eligibility, is cases of grace periods) falls under any of the following:

• No Application

• Zero Income Expired

• Denied (High Income)

• Denied (Incomplete)

• Stated "Does Not Qualify"

• Direct Certification

• Prior-Year Free

• Prior-Year Reduced

• Free (Special Circumstances) - severe need, homeless, Indian reservation, migrant, per administrator, runaway, institutionalized, Headstart free, Headstart reduced, disaster relief, foster child, declining benefits

• Reduced (Special Circumstances) - severe need, homeless, Indian reservation, migrant, per administrator, runaway, institutionalized, Headstart free, Headstart reduced, disaster relief, foster child, declining benefits

• Denied (Special Circumstances - severe need, homeless, Indian reservation, migrant, per administrator, runaway, institutionalized, Headstart free, Headstart reduced, disaster relief, foster child, declining benefits

• Free (Special School Type) - Provision 2, non-participating, institutionalized

• Reduced (Special School Type) - Provision 2, non-participating, institutionalized

• Paying (Special School Type) - Provision 2, non-participating, institutionalized

• Denied (Application Special Circumstances)

 2. Students are grouped by current application and the application is checked to determine if it is pickable.

An approved application is not pickable if the results are:

• Denied, High Income

• Denied, Incomplete

• Denied, Stated "Does Not Qualify"

For each pickable application where at least one (1) student is pickable and not currently selected for verification, a random number is assigned. The algorithm for generating this random number is based on a hash-function of the resulting string of concatenation of the first student's globally-unique identifier and the aforementioned random seed. The application is marked as being error-prone or non-error-prone, being a case-number application or non-case number application, and whether or not someone on the application was picked for verification last year. The list is sorted by the random number and the sample set picked based on that sort-order.

First, if case number applications are needed (if a focused sample is chosen), an attempt is made to find the requested number of case-number applications where no one was picked for verification last year. If more case-number applications are needed, the system attempts again to find case-number applications, but this time ignoring whether or not anyone was picked for verification last year.

Next, if error-prone applications are needed (if either a Basic or Focused sample is chosen), an attempt is made to find the requested number of error-prone applications where no one was picked for verification last year. If more error-prone applications are needed, the system again attempts to find error-prone application, but this time ignoring whether or not anyone anyone was picked for verification last year. If there are still more error-prone applications needed, the system looks at all pickable applications (whether case-number, error-prone, or neither) and chooses from those that were not picked for verification last year. Finally, if there are still not enough picked to meet the requested number of error-prone applications, the system looks at all pickable applications ignoring whether or not anyone was picked for verification last year.

Lastly, if random applications are needed (the Random sample is chosen), the system looks at all pickable applications (whether case-number, error-prone, or neither) and chooses from those that were not picked for verification last year. If there are not enough picked to meet the requested number of random applications, the system looks at all pickable applications ignoring whether or not anyone was picked for verification last year.

If any of the requested number of applications could not be met (case-number, error-prone, or random), a warning message is generated and the user is prompted whether to accept the sample anyway or to cancel the sample generation process.