How can your simulator provide better results than a live trial?
As I mentioned, live trials are often based upon assumptions or theories that have not been rigorously tested. By examining how clients respond and interact with social agencies in a simulated environment, we can gain insight into why a program or policy may not deliver the intended outcome. Our simulator will enable policy makers to identify any assumptions within the framework of a policy that are erroneous or that require modification.
Keep in mind: the goal of the simulation is not to avoid failure but to understand what causes a policy or program to fail specific individuals.
Can you describe your simulator and how it works?
The simulator is designed to emulate the behaviors of both clients and service providers such as social workers. By emulation, I mean that the unique characteristics and decision-making processes of each should be as close to the real thing as possible. And the simu- lated environment where they interact should mirror the network of services available in a specific city. We call this “high fidelity" modeling.
Our approach to emulation is proceeding at two levels: the first is to develop a detailed cognitive model that captures how specific subgroups of social service clients and caseworkers make decisions to reach a goal. The second is more abstract: we’ll be using probabilistic models of decision making to generate expected behaviors of clients and practitioners. With this information, we can run models of various client groups through a simulated network of a city’s social services and see where the process breaks down.
Who will actually use this simulator?
The ultimate users will be policy makers. Many people come together to draft and evaluate social policies and programs. They may include practitioners such as social workers and healthcare workers, social agency administrators, researchers, special interest groups, consultants, as well as elected officials at local, provincial and federal levels of government. Initially we’ll be testing the simulator with social workers in the field.
Mr. Gajderowicz, I understand that your research is instrumental to the success of this project. Before we get into details, can you explain why you chose to focus on the homeless?
Actually, my initial work focuses on single homeless men. They represent the most rapidly growing homeless population in Canada. According to recent studies, 50% of Canada’s homeless are single men, between the ages of 25 to 55. When you consider that 235,000 Canadians qualify as homeless, homeless males become an important demographic to model and study. I’m particularly interested in chronically homeless males. These are men who live on the streets or in places not intended for human habitation for at least four periods within a year. The cost of supporting the chronically homeless is $1.4 billion a year. While chronically homeless men represent around 20% of homeless people, they consume 50% of our social resources due to their need for emergency services, shelters and legal assistance.
Dr. Fox mentioned the concept of “high fidelity modeling” as a requirement for your simulator. What does it mean to create a high fidelity model of a homeless person?
I’ll give you an example: Imagine Jane, a homeless 25-year-old who suffers from anxiety and substance abuse. She tells a caseworker that she wants to enter a detox program. The caseworker knows that addictions are often rooted in mental or physical abuse that occurs during childhood. He learns that Jane is currently in an abusive co-dependent relationship. The client refuses to offer details because she doesn’t like to talk about her personal life with strangers. The caseworker suspects that the client’s abuse and addiction are linked, but he doesn’t know to what degree. Jane insists that she is ready to deal with her addiction, but she is not ready to deal with her partner, on whom she is dependent.
This is a very common scenario that social service providers encounter every day; their clients live unstable, transient lifestyles and deal with many uncertainties on a day-to-day basis It’s the social worker’s job to present progressive steps that are realistic and achievable for a client like Jane.