Across studies there is great variability in reported rates of stuttering recovery. This study examined the impact that different definitions of recovery had on calculation of recovery rates and factors associated with recovery within the same sample of children. Method: Speech samples and parents and child reports of their experiences of stuttering were collected from 38 children who stuttered aged 2–5 years of age (Occassion-1) and again at 9–13 years of age (Occassion-2). Four different criteria for recovery that were developed representing variations in criteria reported in previous research were applied to data from these children. Results: . The majority of the participants (82%) showed very little disfluent speech (<1% syllables stuttered) at Occasion 2. Recovery rate varied greatly depending on the criteria used, ranging from 13.2%–94.7%. Definitions ordered from least to greatest recovery that were (a) parent and clinician report no stuttering and no stuttering observed (13.2 %); (b) ≤1% syllables stuttered; severity rated at ≤1; parent, clinician, and child report recovery (55.3 %); (c) ≤1% syllables stuttered; severity rated at ≤1; parent and clinician report recovery (71.1 %); (d) <3.0 % syllables stuttered (94.7 %). Five participants were considered recovered and two were considered persistent stutters across all criteria. Different factors were associated with recovery from stuttering depending on the criterion used. Conclusion: The concept of recovery from stuttering is complex and estimations of recovery rate are likely to be greatly affected by differences in definitions and measurement across studies. This has a flow-on effect in determining the factors associated with recovery from stuttering.