Public Safety in the Light of City Service Surveys
Using archived data for master's thesis
Harri Myllyniemi 22.9.2004
As I began to plan my master's thesis there were plenty of research subjects to choose from but I thought I would like to
study something of social significance. Citizens' well-being seemed a good choice. I decided to approach the subject with
quantitative methods and survey data. As I preferred extensive, national level data, the best option seemed to use already
existing survey material. My thesis supervisor did not see any problems in using existing data, for him a sound research
design was the main issue.
I had heard that FSD houses interesting datasets so it was time to get acquainted with the archive, formulate my research problem
and start looking for appropriate data. FSD collections had plenty of data on citizen welfare. After familiarising myself with
the archive's collections I formulated my research problem as citizen's safety in their everyday living environment, which I would study with the help of City Service Surveys 1997 and 2001.
Extensive data allow rich and varied analysis
My main data source is the City Service Survey 2001. To assess the stability of certain
trends over time and reliability of the data I compare the 2001 survey to the 1997 survey. The reason for choosing these two datasets was that they contain many background variables relevant for my research problem. Original reports had analysed only a part of the data, so with a new approach and by concentrating to a particular subsection, new findings were possible.
In my thesis I use only a few explanatory variables like, for example, fear of becoming a victim to a crime and whether people feel safe when going out in the municipality. Additional data sources include statistics and various municipal parameters.
Some aspects to consider
Even though it is easier not to have to collect data oneself, using existing data has its own problems. Often the data do not match exactly with the demands of one's research problem. Compromises must be made and the data inevitably influence research problem formulation.
The suitability of the data and particular variables must be carefully considered. One must decide what kind of patterns variables do or do not measure. The data must also be studied critically in terms of validity, reliability, representativeness, skewness and non-response.
At the moment my thesis is still an ongoing project. However, preliminary findings indicate that there are considerable differences between municipalities in relation to, for example, how safe inhabitants feel when outside in the evening. These differences do not seem to be connected in any unambiguous way to particular characteristics of municipalities. In fact, connections seem to be more complex, so there is certainly a lot to analyse.
The data archive has been a great help both when planning and carrying out my research. Research Methods Web Resource (MOTV) has been a help for carrying out analyses. I recommend the services offered by FSD to any student interested in extensive, archived datasets.