• Follow Curiously Persistent on WordPress.com
  • About the blog

    This is the personal blog of Simon Kendrick and covers my interests in media, technology and popular culture. All opinions expressed are my own and may not be representative of past or present employers
  • Subscribe

  • Meta

Considering the collective opinion

Life is not logical. The whole is different to the sum of its parts.

Yet the majority of quantitative research is not structured to reflect this.

It assumes that survey respondents are considered and rational individuals, able to make an informed choice when asked to consider a series of options, or able to adequately communicate their thoughts, opinions or attitudes.

Yet, in most instances this isn’t true

  • The school of behavioural science, currently championed by Rory Sutherland at the IPA, portrays humans as mannequins to be manipulated by an invisible hand, or nudged to predictably irrational conclusions
  • Those who believe Google makes us stupid talk about the rise of distributed memory. We are now more wired to remember a wide range of data superficially, rather than a depth of informed knowledge
  • Our hectic lives preclude a degree of considered choice on non-special occasions, so we blink, “thin-slice” and make gut reactions
  • The paradox of choice and analysis paralysis mean we prefer to make choices on a more limited number of options than are perhaps available – this screening process is by no means optimal
  • The individual is not an island, but part of a community

This latter point in particular has been on my mind recently.

There have been attempts to engineer research by identifying influentials or mavens (link is a pdf) but I’m not aware of any technique that satisfactorily conveys the intertwining impact that our personal relationships and media exposure have on our thoughts and actions.

We might follow Herd behaviour. So, if the media constantly mentions Facebook, it leads people to try it. And if Academy Members hear that the Hurt Locker is the favourite to win best film, then they will vote for it.

But then there doesn’t appear to be a guarantee that this prophecy is self-fulfilling. A Labour victory at the 1992 election was seen as a foregone conclusion, yet people stayed at home or kept their Conservative biases hidden until the day of polling. And the hype surrounding Second Life didn’t convert us all to living vicariously through an online avatar.

So why not? Duncan Watts has worked extensively on this. He argues in favour of randomness, where an initial option gains some traction that is in turn exacerbated as people validate it. This makes some intuitive sense – if I am searching on Amazon, I might initially limit my search to the highest rated items, even though those rating them might have highly different needs for the product to my own.

Is it possible to predict which ideas or concepts will eventually succeed? I know Brainjuicer have had some success with their predictive markets, but it appears to be an area open for innovation.

Would we need to run full simulations (either controlled experiments, or computer models) to judge which environmental factors will have the largest influence?

Or would it be possible to infer a “herd multiplier” to frame a concept in the likelihood of it being adopted by a community or the media?

Or could we go back and question people on their expected behaviour in the context of what their actual behaviour was, in order to get them to post-rationalise the differences?

Or, will this always be a problem with quantitative research, and instead we should look to smaller sample groups of people to test how ideas and preferences disseminate and iterate through a group of likeminded people?

I don’t know the answer. But it is certainly something worth thinking about.

sk

Image credit: http://www.flickr.com/photos/28191556@N02/3160824946/

Mark Earls – From “me” to “we”

Thanks to Mat kindly donating his ticket, I was able to go and see Mark Earls give a seminar entitled From “me” to “we” at the Royal Society.

herd by mark earlsRather shamefully, I am still yet to read Herd – the book (and associated research) on which the talk was based. This is despite regularly reading the Herd blog and even having a copy in the Essential library. As I said, shameful.

Despite this, I think I was the target audience. Along with a Q&A only notable for the rather aggressive questioning of a lady accusing Mark of ignoring “the female perspective”,  the session offered a fairly gentle precis of the book’s central theory which, if I had read it, I would of course have been familiar with.

The talk

A tenet of the book is that we’re bad at changing other people’s behaviour. To highlight this, Mark recalled a few statistics from his research:

  • Only 10% of new products survive longer than 12 months
  • Only 30% of change management programmes begin to achieve their aims
  • Mergers & Acquisitions lessen shareholder value two thirds of the time
  • No government initiative has created demonstrable and sustainable change

This is particularly worrying because behavioural change comes before attitude change – our thinking comes after the fact. We (post)rationalise rather than act rationally.

Therefore, in order to change attitudes, we need to change behaviour. And to be able to do this, we need to understand who we are. Only then can we can create solutions that work.

The Herd thesis draws upon the Asian culture of believing that humans are naturally social. We are fundamentally social with only a bit of independence, not vice versa.

Although it doesn’t sound particularly controversial, this thinking does run contrary to some well established tenets of both marketing and social theory.

According to Mark, thinking is much less important in human life than it seems. He likens us and thinking to a cat in water – we can do it if we have to, but we don’t particularly like it.

This is because it is easier to follow than think. We know our judgement is fallible and so we outsource the decision by following the crowd. But while this may work in some situations – many illustrated by James Surowieki – it is also arguably a contributing factor to the financial crisis, as financial institutions copied one another without comprehending the implications.

We therefore need to design our theories and tools to accommodate this social behaviour. It is much more rewarding to understand how social norms are created and perpetuated than it is to work on the assumption of cogito ergo sum.

Some initial thoughts

While brief, the talk certainly conveyed the need for me to read the book fully. Perhaps then some of my questions regarding the theory will be answered.

In particular, I’m interested in knowing where movements originate and whether this herd behaviour can be predicted.

For all the sheep, there must be a shepherd somewhere. Are these shepherds always designated as such – the almost mythical influentials – or do we alternate between thinking and following?

Rarely are our choices as clear cut as choosing whether to join the corner of the party where people are talking rather than the one where people are sitting in silence. Instead we have multiple choices and herds – how do we choose?

Is it a level of proximity? In the Battle of Britpop, Northerners sided with Oasis and Southerners with Blur? However, I’m from the Midlands, so was my choice one of the rare occurrences of rational choice (which would make a rather unconvincing deus ex machina) or is it purely random?

If random, then the work of Duncan Watts becomes pertinent. His modelling has suggested that in situations where groups vote up and down their favourite songs, there is no objective winner. Different simulations create different patterns. Purely random.

This creates difficulties for researchers as we like our statistical certainty. We like to have a set answer that we can post-hoc explain given the evidence. Duncan Watts’ research would suggest that research tools that build in mass opinion – such as crowdsourced tagging or wikis – are effectively meaningless. Rather than ultimately deviate towards a “correct” answer, they simply reflect the random order of participation and interaction.

Can mass behaviour be effectively incorporated into a research programme? I’ll report back with some thoughts once I’ve read the book

sk

We’re bad at changing other people’s behaviour

Only 10% of new products survive longer than 12 months

30% of change management programmes begin to achieve their aims

Mergers & Acquisitions lessen shareholder value 2/3 of time (pwc)

No government initiative has created demonstrable and sustainable change

Behavioural change comes before attitude change – thinking comes after the fact

In order to change attitudes, change behaviour

We need to understand who we are so we can create solutions

More rationalising than rational

Cognitive outsourcing – memory is a distributed function so only remember slivers

We are fundamentally social with a bit of independence, not vice versa

Asian culture is inherently social

Gandhi said that humans are a necessarily interconnected species

Thinking is much less important in human life than it seems

“lazy mind hypothesis”

We can think independently, we just don’t like it – like a cat to water

Behave according to other people’s actions e.g. go to busy shops

We know our own judgement is fallible so “I’ll have what she’s having” – wisdom of crowds or financial crisis

Leads to social norms

Need to design our theories and tools to accommodate social behaviour

Genesis random – Duncan watts

Is it proximity that leads us to follow a herd, or example of using rationally weighing up the pros and cons

Herds originate from somewhere – must be a leader. Are these leaders the same in each situation, or are we all capable of being shepherds

Research application – crowdsource answers. But random – no statistical certainty as only one situation

Wikis to collate group opinion?

Reblog this post [with Zemanta]