This study used negative binomial regression to investigate whether exposure to novice teachers and risk for identification for special education predicted suspension rates. Data from the 2009-2010 Civil Rights Data Collection (CRDC) was used. The sample was comprised of 72,168 schools from nearly 7,000 school districts from nearly every state. Identification as having emotional disturbance and specific learning disabilities were found to predict an increase in suspension rates for some subgroups across some school levels. Conversely, identification as being autistic was found to predict a decrease in suspension rates for some subgroups across some school levels. Policy implications are discussed.