For each U.S. induction program, researchers twice observed and interviewed 5-10 beginning in each of two years, and observed mentor-mentee meetings as possible. We also contacted some beginning teachers for brief phone conversations during the summer. We enlisted a mixture of first- and second-year teachers. Interviews also were conducted with mentors, program leaders, event facilitators and some building administrators. Beginning teachers in the U.S. all were volunteers, recruited by the researchers and selected to constitute a range of experiences in the program (e.g., low and high participation) and teaching contexts (urban/rural, teaching different mathematics courses, etc.). While biases introduced by this non-random sampling are not clear, researchers often were able to observe all beginning teachers during group induction events and informally consider how study subjects were generally representative of the population. Researchers did not share data with program providers in real time, in order to limit any potential influence they might have on participants' behavior and disclosures within the study interviews. Flexible observation and interview protocols (each require a class period) emphasized mathematics- or science-specific aspects of being a beginning teacher and, therefore, could not cover the entire terrain of a beginning teachers' experience. Researchers had no staff or consultant role in any induction program, or any prior relationships with the study subjects. Senior researchers had extensive prior work in qualitative research and every researcher had expertise in one or more field: teacher learning, teacher induction, mathematics education or science education.
Researchers wrote observation notes, and recorded and transcribed interviews. Data from a precursor study to the MSP RETA were analyzed through a grounded-theory approach where multiple researchers examined data from study subjects, triangulating the multiple sources of data about each subject. Teams of researchers further analyzed data across all study subjects to write ongoing analytical memos that yielded emerging themes and examined evidence for rival hypotheses. Data in later school visits was shaped to further investigate evolving findings, questions and competing hypotheses. For data in the MSP RETA study, researchers are coding all observations and interviews against a set of over 80 topics/codes in the Hyperesearch software for qualitative analysis. The software permits analysts to extract from the entire data record every instance of a topic discussed or observed. Teams of researchers with expertise relevant to a code are analyzing data across all study subjects. However, the researcher who collected data from a study subject, for each visit with them, also conducts a holistic analysis of all data about the individual by addressing 11 analytic questions developed by the study team.