Here we summarize the ethical considerations associated with our questionnaire survey on the bio-psycho-social factors underlying convicted child maltreatment cases in Japan. The purpose of this research is to obtain background information and support needs directly from the parents and cohabitants who conducted child maltreatment, in order to develop preventive and supportive measures for families at risk of severe child abuse. We have been conducting a survey research since 2016, with caregivers (or cohabitants with children) who were incarcerated for their child abuse conviction. Scientific research on human participants with criminal records requires thorough ethical considerations, such as informed consent and privacy protection, as much as clinical research that entail mental disorders does. This research was approved by our Ethics Committee, and was only conducted at facilities that were approved by both the Correction Bureau of Ministry of Justice and the correctional institution itself to be involved, after our extensive explanation and communication about the purpose, significance, and the contents of our research. We collected information of child abuse cases, in which the perpetrator was sentenced to prison and whose name was reported in the media with real name. We requested participation via postal mail, and those who replied gave written consent. Control parents who have some experience of parenting were collected by advertisements. The survey consisted of more than 400 questions, which were constructed as needed from questions regarding their life history, the Beck Depression Inventory (BDI), Adverse Childhood Experiences (ACE), the SCID-II personality questionnaire, traumatic experiences, Parental Stress Index (short form), or about stress factors around the time of the crime. These were performed via three rounds of postal mail between RIKEN and the participants in penal institutions, starting with the least stressful of the questions. Unclear answers were reconfirmed with the participants for further clarity. To balance the scientific validity and the risk for participants, cautions were taken such that i) questions about the facts of illegal acts were replaced with more cognitive questions, ii) indications were placed to stop answering the survey whenever the participant experienced psychological stress by the questionnaire, iii) we remained neutral for the incident and its judgement as an independent research organization. The survey is still ongoing, but during the four years, 36 perpetrators provided written consents and 31 successfully finished the four rounds of mail survey. The number of control parents was 178. Surprisingly, we received multiple positive remarks from the convicted participants for attending the survey at the end of the communication. The processes, findings, and remaining challenges of conducting this sort of research are further discussed.
 Authors' abstract