SENG3011

🐶  8.2 - Influencing Behaviour

In this lecture

  • Theory of reinforcement
  • Types of reinforcement
  • Delivering reinforcement
  • Incentives, rewards and punishments

Examples of classic 'incentives'

  • In a business context:
    • Wages/salaries
    • Bonus for reaching annual targets
    • Performance appraisal (quarterly review)
    • Employee of the month
    • Disciplinary procedures
  • In a university context:
    • Marks
    • Activity is fun
  • At home (for kids)
    • Treat foods
    • Screen time

Recap: Why do people behave the way they do?

  • The individual
    • Personality 
    • Capability
  • The organisation
  • The situation
  • People behave how they have been conditioned to behave
  • Everyone's behaviour makes sense from their point of view

Behavioural consequences

  • Behaviour is a function of its consequences
    • People do what they do because of what happens to them when they do it
    • Performance is what about happens every day
  • Behavioural consequences - those things and events which follow a behaviour and change the probability that the behaviour will be repeated in the future 
  • Most of the behavioural consequences we encounter we don't think about (e.g. turning on a light, typing on a keyboard)

How to influence behaviour

  • How to make people stop doing things:
    • Identify behaviours that are producing the poor outcome and arrange consequences that will stop them
  • How to make people start doing things:
    • Identify behaviours that are producing desirable outcomes and arrange consequences that will positively reinforce them

Four ways to change behaviour

ABCs of Human Behaviour

  • Antecedent (the situation preceding the action)
  • Behaviour (the action itself)
  • Consequence (the outcome of the action) - reinforcer
  • AKA - Given, When, Then
  • Difference between can't do and won't do
    • If someone did something correctly in the recent past, but isn't doing it anymore - it's probably a won't do

Three dimensions to a reinforcer

  • Positive / negative - is the consequence positive or negative from the perspective of the performer?
  • Immediate / future - does the consequence occur as the behaviour is happening (immediate) or some time later (future)
  • Certain / uncertain - what is the probability that the performer will experience the consequence?

Example: Recycling

Example: Writing tests for your code in SENG3011

Can we spot a pattern of consequences types?

Trust

  • Trust is doing (consequence) what you say you are going to do (antecedent)
  • Tell someone something's going to happen => It happens => +ve reinforcement
  • Tell someone something's going to happen => It doesn't happen => -ve reinforcement
  • Be careful of over-promising

The price of negative reinforcement

  • Positive reinforcement incentivises maximum performance, while negative reinforcement gets a level of performance that is just enough to get by to escape the unpleasant consequence
  • Indicators that negative reinforcement is present:
    • People get things done just in time
    • Most of the work happens in the last few days (J-curve)
    • Negative talk
    • Performance goes flat after reaching a goal
    • When you remove a performance requirement and performance drops

Negative reinforcement has its place

  • Negative reinforcement serves us well when all we need is compliance/minimum performance
  • If you can't find anything to positively reinforce, then the person is probably in need of some negative reinforcement
  • Negative reinforcement needs to be paired with positive reinforcement of improvement
  • B. Franklin "Write people's accomplishments in stone and their faults in the sand"

The tricky thing about negative reinforcement

  • Negative reinforcement is more likely to provide us (the reinforcers) with a PIC than positive reinforcement
  • We need to wait until the next time a behaviour occurs to see if the reinforcement work (PFU)

Positive reinforcement

  • Discretionary effort - the level of effort people could give if they wanted to, but is beyond what is required
  • Natural - e.g. turning on a light
  • Created - must be added by a person
    • Social - involves doing or saying something to another person and has trophy value (e.g. a compliment)
    • Tangible - an object and has salvage value (e.g. a toaster)
    • All tangible reinforcers should be paired with social reinforcement
  • Find reinforcers for people: ask, try, observe
    • Different people will have different reinforcers 

Positive reinforcement

  • Grandma's Law - if you eat your veggies, you can have dessert
  • Pair actions that aren't reinforcing with ones that are
  • Sources of positive reinforcement:
    • Work-related reinforcement - when we arrange a task so that reinforcement is automatically associated with the task (e.g. green CI ticks, tests passing)
    • Peer-related - from peers at work - in best position to deliver PICs, can observe performance more closely and more often
    • Management-related 

Decreasing behaviour

  • Punishment + Penalty
  • Don't shoot the messenger
  • Delayed punishment is no more effective than delayed reinforcement
  • Punished behaviours should be paired with positively reinforced replacement behaviours
  • Recovery - old behaviour resurfaces

Extinction

  • Doing nothing changes behaviour
  • Extinction - withholding or not delivering reinforcement for previously reinforced behaviour
  • "Just ignore it and it'll go away"
  • Signs of extinction
    • Extinction burst (e.g. lift button)
    • Erratic/emotional behaviour
    • Resurgence

Errors in delivering reinforcement

  • Perception - does the performer find it reinforcing?
  • Contingency - must be if and only if (desired behaviour) then (reinforcement)
  • Delay - optimal reinforcement is immediate
  • Frequency - has to happen at a frequent rate (annual, quarterly, monthly isn't frequent enough)
  • 5 to 1 rule - 1 negative remark equals 5 positive remarks
  • ... "but" ...
  • Shit sandwich - avoid pairing positive reinforcement with punishment
    • Shit sandwich is still good for feedback

Let's revisit our incentive examples

  • In a business context:
    • Wages/salaries
    • Bonus for reaching annual targets
    • Performance appraisal (quarterly review)
    • Employee of the month
    • Disciplinary procedures
  • In a university context:
    • Marks
    • Activity is fun
  • At home (for kids)
    • Treat foods
    • Screen time

How has everything changed?

  • "Why do we have to do all this reinforcement stuff today? We didn't use to have to do it, and we got along OK"
  • We are conditioned to behave the way we do
  • How many positive reinforcers do you get a minute playing a computer game/watching YouTube?

Further Reading