Outcomes are important for data driven decision making. Macro-level outcomes are the “big picture” results the majority of people experience. Policies and practices are largely driven by macro-level outcomes.
Micro-level outcomes refer to the outcomes or results a person experiences based on their unique set of individual and social circumstances. Micro-level responses are those individual decisions made every day by professionals in the social services and juvenile justice fields; the daily judgement calls made by judges, caseworkers, juvenile probation officers, youth care workers, police officers, etc. These individuals make decisions based on best practice guidelines and professional judgement many times throughout the course of their workday.
If these micro-level decisions are mostly made by following best practice guidelines, then why are macro-level outcomes, such as disproportionate minority contact in the justice system, under such scrutiny?
The reason for this misalignment is because macro-level data is sometimes used to inform micro-level decisions. In the juvenile and criminal justice fields, a frequently used macro-level outcome is recidivism or re-arrest rates, relating to a specific program or agency. These outcomes are important because they show an overall social impact, often either positive or negative, which then informs decision making.
However, the mismatch in the decision making process happens when macro-level data is used to inform micro-level decisions about an individual. That mismatch is often times why macro-level outcomes, such as disproportionate minority contact, are so challenging to improve.
Let’s take a look at an example of how macro-level data is used to inform micro-level decisions in the juvenile justice field:
Johnny is a 15 year old youth involved in the juvenile justice system who has a long history of offenses and has been identified as needing treatment outside of his community and home. He participated in many community-based services, most of which were unsuccessful. A risk assessment used by the probation department identified Johnny as a “high risk” to reoffend.
Johnny’s juvenile probation officer and his supervisor meet to discuss where they should place Johnny for treatment.
Probation Officer: I think we should make referrals to Program A and Program B. Program A has a treatment curriculum to meet his high need areas and he can play sports if he goes there. Sports might motivate him to do well and complete the program. Program B will also work with him on his high need areas and Jimmy just left there last week and he seemed to have really done well and learned a lot.
Supervisor: Okay, according to their data, Program A is reporting 75% of youth who go to their program successfully complete the program, and Program B is reporting that only 50% of their youth successfully complete the program.
Probation Officer: Oh wow, we can’t afford to send him to Program B with only a 50% chance of completing the program. I think we should just make the referral to Program A.
Supervisor: I agree, make the referral to Program A. Don’t bother with a referral to Program B if there’s a 50% chance he’s going to fail there anyway.
So in this example there are two macro-level outcomes:
- Program A has a successful completion rate of 75%
- Program B has a successful completion rate of 50%
The logical side of your brain would agree with this probation officer and supervisor. Send Johnny to Program A!
We should consider the following questions:
- What about the 25% of youth who went to Program A that were unsuccessful?
- What about the 50% of youth who went to Program B that were unsuccessful?
- What did they have in common?
- Why were they unsuccessful?
- Where would those youth have had a better chance of success?
- And what about Johnny, whose future could be determined based upon macro-level data?
- What if Johnny falls in the 50% of youth who do well at Program B?
- What if he falls in the 25% of youth that will fail in Program A?
- Is there a better way we could be making these types of decisions?
DECISION MAKING TOOL
FirstMatch® is an innovate solution that uses artificial intelligence and machine learning to address the unique challenges of operationalizing both micro and macro-level data. The software application uses predictive analytics to generate predictions about the likelihood an individual, rather than a program, would experience various outcomes.
Outcome predictions can be made for treatment program completion, remain-out-of-care, permanency maintained, re-hospitalization, risk reduction, recidivism reduction, and other pertinent outcomes of importance to the treatment of the individual.
If utilized, FirstMatch might have revealed that Johnny, based on his unique set of circumstances, would have had a 40% chance of completing Program A and an 85% chance of completing Program B.
By analyzing the individual with all their unique circumstances, and selecting the best course of action for that particular individual, it’s likely that person would experience a more successful outcome therefore, in time, impacting macro-level outcomes.
Now, which program seems like a better option for Johnny?