Ask better Questions: Algorithms for everyone!

June 27, 2014 by
Filed under: Innovation, Social Economy 

 

I finish my series of 5 community rules, 4 principles of engagement3 approaches that work, 2 ½ change caveats, with One Ask.

My one ask is: Ask Better Questions.

1. Local-level practitioners should be using our 5 community rules.

For example, when you go to see a GP, they should be asking about your social connections. Do you have any? Can they link you to local groups? Do the connections you have already have the capacities to help you?

2. When local authorities make service allocation decisions, they should be informed by a holistic understanding of impact.

 

This could be using something like our 5 community rules to ensure that new policies don’t negatively affect someone’s social contact availability, doesn’t harm the already vulnerable, makes people feel more part of a community and not less. So, does closing down a post office remove an elderly person’s only source of social contact; does moving a young family to an area of cheaper rent remove the social support that make being a single working mum feasible; does driving vans with “Go Home or Face Arrest” achieve anything other than making people more divided than before?

3. Everyone else should think about sharing their own ‘best practice questions’.

Introducing… Algorithms for Good!

Algorithms for Good?

Algorithms are everywhere. You may not be able to see them, can’t touch them, and don’t know what they taste like; but they see you and likely know both what you’d like to touch/own/buy and what your tastes are. Go onto Amazon and it will recommend purchases based on your browsing history. Go onto twitter and it’ll recommend people based on who you’re already connected to. Ask an investment banker a questions and it’ll be the algorithms doing all of the answers.

My question is: what would ‘Algorithms for Good’ look like?

Algorithms are basically rules of thumb, scaled up; they are “a process or a set of rules to be followed in calculations or other problem-solving operations, especially by a computer.” An algorithm for making tea might be something like this:

via http://www.technologystudent.com

Now, it would be ridiculous to have to function like this, day-to-day, but if we all suffered collective tea-making amnesia, this algorithm might be a god-send.

My offer:

So what questions might we ask? Social Mirror combines tablet computers and algorithms to issue ‘social prescriptions’ in a digital form. It asks a series of questions about someone’s wellbeing, social connections, health and interests and then links them to activities and groups that might help or be of interest, thus addressing poor health, mental wellbeing and isolation.

Social Mirror came about we know that social prescribing can be very positive for people, but that its use depended on the practitioners. Does the local GP think it’s a worthwhile endeavour? Does the local GP know what is happening locally? We decided to put our local knowledge and connections in Knowle West to good use, combining them with the RSA’s specialist knowledge in wellbeing science and social network analysis.

A retired gentleman who is isolated (has no social connections and/or says he is lonely) and has an interest in being active might be issued a prescription to a walking group:

“I can’t say enough about it because it has changed my life. If I hadn’t done it I wouldn’t have known about these walking groups. After I retired I felt like a recluse, three days a week I didn’t go out of the flat. I’ve now lost a stone in weight, I can talk to people quite freely which I didn’t before. I’ve stopped drinking alcohol – I don’t need it to help me sleep as the walks tire me out.”

A young mother who is new to the area (so comes up with a low community score), and does not know many people locally (so has very sparse connections), might be prescribed a children’s group:

“Social Mirror has made a massive impact in my life because when I moved here I had nobody and nothing. Going to groups through Social Mirror started the ball rolling – I’ve been going to groups for my children and for myself, I’ve made friends, and I know the area better. My life is a lot happier and more content now and I don’t feel so lonely.”

My Ask:

What happens when these algorithms are allowed to become bigger than us and our knowledge?  So far, we have coded based on knowledge and experience that is either in-house at the RSA, or easily accessible within our contacts. We based the questions and algorithms on our research with 3000 people across England, analysis of the ONS’s and the ESS wellbeing data-sets and what we know about social network and wellbeing theory (quite a lot!). We asked friends at the NEF wellbeing team for specialist advice, and did a LOT of testing with local people and community and health practitioners.

But none of us are pregnancy specialists. I don’t know much about being a teenage boy. We haven’t had to juggle being a full-time carer and a full-time mum. What happens if we had the resources to open our approach to health, social care and life experts? For example, pregnancy (and world cups football games!) are a major trigger for first-time domestic violence; pregnancy is often the first time that women come into contact with the state as adults: what are three questions I should be asking to make sure everything is ok?

This Social Mirror approach could work across so many fields: ask questions, apply simple algorithms, issue appropriate prescriptions. You take specialist knowledge, pare it down to a usable mathematical formula and apply it locally. As with everything, it would obviously be better to have the expert in the room, to have time for all the conversations, but until we invent cloning and time-travel then maybe something like this will have to do: Social Mirror becomes the central repository for the ‘pure’ knowledge, it then becomes applied locally on a case-by-case basis.

I’ll show you mine, please show me yours.

I argue it’s good to have an offer: these are the questions that I would ask – of a policy, of a person registering at their GPs – based on the work we have done with 3000 people in deprived areas of the UK. What would your burning questions be?

my questions_________________________________________________________________________________________________

Gaia Marcus is a Senior Researcher on the RSA Connected Communities project.

She is an Edgeryder and an UnMonk advisor, founded the RSA Social Mirror project and is ¼ of the ThoughtMenu.

This Summer she will be cycling 1000km in memory of our RSA colleague Dr Emma Lindley and in aid of Mind

You can find her on twitter: @la_gaia

Comments

  • Johan

    The example with the tea-making is not very good. Why is there even a ‘No’ option for the ‘Turn off tap when full’?
    It should have been:

    Add water to kettle <–
    | |
    V | (No)
    Is Kettle Filled? ———
    | (Yes)
    V
    (next stage)

    And the arrows that points to a previous stage, points to another arrow. Which makes it confusing. They should point to the stage that needs to be done instead of the arrow coming from that stage.
    This was just an awful example. With unclear directions when something is not going as planned.

    • Johan

      Add water to kettle <—
      ………|…………………….. |
      ……..V……………………. | (No)
      Is Kettle Filled? ———
      ………| (Yes)
      ……..V
      (next stage)

      • Gaia Marcus

        If you have a better image, send over and I will replace.

    • Gaia Marcus

      Hi Johan, Thanks for your comment.
      How do you feel about the generic argument, beyond whether the flow diagram represents a series of ‘if’ function algorithms sufficiently for a lay audience?