In answering this question, I will take self-reported data and victimization data to be the same thing. The reason for this is that victimization data is data that is obtained by asking people about whether they have been the victims of any crimes. This is, therefore, a self-reported type of data.
The major strength of official data on juvenile crime is that they are objectively true in an important way. When an official record is made of a crime, it means that a crime has clearly been committed. There is no need to wonder if the report is accurate. A second strength of such data is that it is separated out into consistent categories. This makes it easier for a researcher to identify the nature of the crime that was committed. The major weakness of this sort of data, however, is that it may not necessarily cover all of the crimes that are actually committed. There may be many crimes that are committed but which are not reported to authorities. This could happen, for example, because the victim does not trust the authorities or because he or she fears retribution for going to those authorities.
This is where the strength of self-reported victimization data arises. This is data that has been collected by surveying people. It allows them to honestly report whether they have been victims of a crime and to do so in an anonymous way. Therefore, these reports are arguably more comprehensive and more truly representative of reality than official data are. However, self-reported data are only accurate if the people who are surveyed are honest and if they have a good concept of what a crime really is. If they choose not to tell the truth, or if their definition of a crime is different from the official definition, the data can be less than accurate.
Thus, each of these types of data has both strengths and weaknesses.