Data Collection

In March 2018, two of the authors independently conducted electronic searches of all 50 state Departments of Education websites to obtain policy documents authorizing APs. We omitted programs that included reciprocity agreements with other states or required National Board Certification, as both require a traditional license. We also recorded data from policy documents as well as extensive notes to explain complexities and idiosyncrasies in the policies.

Data Coding

We developed a coding manual to ensure that we coded policies consistently. Working independently, two of the authors coded programs from 25 states each on several features, including program title, eligibility, program and testing requirements, and time limitations for program completion. We entered these data into two separate Excel databases. To estimate inter-rater reliability, we conducted a second round of coding in which we exchanged states and, using the coding manual, coded them independently. Inter-rater reliability was 90%. With discussion, we reconciled all differences and then merged the two databases to create a single file comprising all authorized alternative route policies across the 50 states.

Data Vetting

Between March and July of 2018, the authors endeavored to contact all 50 states’ departments of education by email, telephone, or both, to verify the accuracy of the information in the database. We first emailed DOE licensure and certification personnel seeking assistance. To non-responders, we sent a second email request in April. If a response was still not received after the second email message, we phoned the DOE contact to ask for assistance with vetting. Ultimately, via email and telephone, we affirmed the correctness (or corrected misinformation) for 35 states. When unable to contact DOE personnel, we also reached out to university colleagues in those states and, with them, vetted three additional states. For these 38 contacts, we documented all communications and took extensive notes as evidence on which to base our decisions. We were unsuccessful in confirming the accuracy and currency of policies for the remaining 12 states.

The authors completed the vetting process through the original process of data collection and verification. We recognize that state licensure policies change, but due to the time-consuming nature of vetting policy data, we are unable to make additions or changes to this data.

Program Classification Definitions

After vetting, we divided states equally and independently sorted policies into conceptual groups based on commonalities in their titles, eligibility, or program specifications. We met to discuss the sets of categories, reconcile differences in our categorizations, and come to agreement about classification. After accounting for overlapping categories and discussing disagreements in the grouping of programs, we arrived at the eight final mutually exclusive conceptual models displayed in the “Type” column. Finally, we individually grouped the program policies in our assigned states into eight categories: