Compared to the rest of the European countries Denmark has the second highest rate of burglary. This makes burglary a well-discussed phenomenon in the Danish media, which seems to cause a high sense of insecurity among Danish citizens regarding burglary. Because of this the Danish Police Force, face a big problem - but how can it be solved? First of all the solution must be to understand burglary as a phenomenon.

"One out of a hundred know about it" is a quantitative document analysis of the quality of police data concerning burglary. The aim of the master thesis is to describe the possibilities to understand burglary from statistics created from police records and therefore the focus is on the quality of the data.

The study is based on 495 records produced in the police district East Jutland in the period May 2014 to April 2015. Because of the use of a stratified random sample we find it possible to generalize the findings of the master thesis to the records dealing with burglary produced by the Police of East Jutland. However, it can be discussed whether it is possible to generalize the results out to the remaining district of the police in Denmark.

Data quality is a multi-dimensional phenomenon why it is necessary to understand and examine several aspects in the desire to understand data quality as a whole. The data quality in this paper is examined by the dimensions accuracy, completeness, consistency, usefulness and relevance.

Because of the unexplored nature of the area concerning quality of the police’s data we find it necessary to deepen our understanding of the field. As a result of this we conduct interviews with two experts.

The results of the quantitative document analysis show that the data quality of the police records sufferers form great validity problems, especially due to the accuracy and completeness of the information in the records. The validity problems are caused by the many discrepancies between the parts of the records which statistics can be generated from and the parts of the records where the burglary is described in depth. Furthermore the data suffers from completeness issues as a result of missing values and relevant information being incompletely stated. In addition the results show that these issues are not systematic in relation to the police officer or caseworker who produce the police records or in relation to the victim of the burglary.

Regarding the estimated accuracy and completeness of the data we find that the usefulness of the system used by the police to generate the statistics causes problems. This as a consequence of the lack of implementation of the statistical tool that aims to systematize and standardize the search criteria in a record.

Furthermore we find that much relevant information can not be generated into statistics, though the information is stated as important by criminological theories concerning routine activity and situational crime prevention. This being information about the characteristics of the property.

All of the above causes issues regarding crime analysis based on the data generated from the burglary records - most of all because of the great validity problems.