Teachers new to the profession may face various challenges and struggle with pedagogy and classroom management. They tend to be less effective in boosting student learning than their more experienced colleagues (Murnane & Phillips, 1981; Raymond, Fletcher, & Luque, 2001; Rivkin, Hanusheck, & Kain, 2001). Since the early 1980s, there has been an increasing recognition of the importance of providing induction support in forms of mentoring programs, workshops, orientation seminars, collaboration opportunities, and other support systems to new teachers in their initial years of teaching (Furtwengler, 1995). At the present time, 27 states require some forms of induction or mentoring support for new teachers, 22 states mandate completion of or participation in an induction program for advanced teaching certification, and 17 states provide dedicated funding for teacher induction. While the general goal of teacher induction is to transform a student of teaching into a competent teacher of students, many evaluations in the past have focused on program impacts on novice teacher retention and professional well-being. Only a few studies have attended to instructional improvement as outcomes (see reviews by Ingersoll & Strong, 2011; Strong, 2009; Wang, Odell, & Schwille, 2008). Most studies (Davis & Higdon, 2008; Evertson & Smithey, 2000; Stanulis & Floden, 2009; Thompson, Paek, Goe, & Ponte, 2004) have suggested that more intensive mentoring and support from university-trained mentors might be associated with a higher rate of using effective instructional practices among new teachers. Yet one study (Roehrig, Bohn, Turner, & Pressley, 2008) reported that new teachers regardless of induction intensity declined in their use of effective teaching practices over the first year. These evaluations have been mostly non-experimental or quasi-experimental with a relatively small sample size. In contrast, a large-scale randomized study funded by the U.S. Department of Education and conducted by a research team from Mathematica Policy Research (Glazerman et. al, 2010) compared two prominent Comprehensive Teacher Induction (CTI) programs with standard district or school support for more than one thousand new teachers. Although teachers in the treatment group experienced more intensive, structured, and sequenced mentoring activities from trained external mentors, they exhibited surprisingly similar teaching practices as those in the control group in the spring of the first year such that a zero effect of the CTI programs was concluded. Reanalyzing data from the comprehensive teacher induction study, the authors aimed to unpack the zero effect of the CTI programs on teaching practices by closely examining the content and activities of mentoring as potential mediators of the induction program effects on teaching practices. The content of mentoring includes teaching planning and preparation, management of classroom environment, instructional content and pedagogy, and professional responsibilities. Key activities for mentees include keeping record and analysis of teaching and student learning, working with a study group of teachers, observing other teachers' teaching, and meeting with local instructional leaders. The following questions were asked: (1) Did treatment teachers and control teachers have different experiences with mentoring content and activities? (2) Did the differences in mentoring experiences mediate the program effect on teaching practices? (3) Was receiving mentoring from external mentors in the CTI programs as effective as receiving mentoring from home-based mentors under the control condition? Preliminary analysis indicated that treatment teachers and control teachers had different experiences with mentoring content and activities. Clearly, beginning teachers assigned to the CTI programs tend to receive a higher dosage of induction content and a higher intensity of mentoring activities. Therefore, we can rule out the second explanation for the zero effect of the CTI programs given that the treatment teachers displayed an equal or higher rate of participation than did the control teachers. The authors did note that a higher level of participation rate in the treatment group apparently did not lead to superiority in teaching practices in comparison with the control group. One would wonder, had the treatment teachers participated in the CTI programs at a lower rate that becomes equal to the control teachers' participation rate in their local induction programs, whether the teaching practices of the treatment group would become inferior to that of the control group. Tables are appended.