On the path towards a merged psychological-network theory of innovation diffusion through online communities

I’m trying to converge two theories in my last empirical study – an exercise that follows the diffusion of voice through the Second Life social networks (interpersonal pathways of connected Friends) from August 2007-April 2008. I’ve got three theories I’m relying on here: the Theory of Planned Behaviour (Ajzen, 1985), the Technology Acceptance Model (Davis, 1989) and Diffusion Theory (Rogers, 1962).

TPB, at its most basic, argues that behaviour is the outcome of intent, which is generated through positive attitudes towards an innovation (including, in this study, the perception that the innovation adds value (socially) and that it perceived as better than what has come before, namely text-based communication), the belief that one can perform the behaviour (behavioural efficacy – i.e., that they’re not limited tchnologically, environmentally or physically) and the expected outcomes (based, in this study, on the social norms of the network). Here’s a piecemeal graph:

TAM, at its most basic, can be viewed as an adaptation of TPB, but specifically for technological systems (and even more specifically, for information systems). It’s been applied to a whole variety of technologies, from the World Wide Web to voice mail. It says that technology is accepted because it is perceived as useful and it is considered easy to use. TAM’s so like the TPB (minus intent and social norms) that I won’t show it off here. Why it’s important, though, is beacuse the perceived usefulness in the case of voice is that it is what’s called an interdependent innovation: other people must use it for it to be viewed as useful, and therefore for it to propagate through the system.

Finally, Diffusion Theory, at its most basic, says that network features – like position in the network and exposure to an innovation from direct network neighbours – contribute to adoption. There are variants: purely structural theorists say that all that’s needed for adoption is the right network position and the appropriate level of exposure to hit an individual’s threshold. Network analysts who veer on the sociological or phenomenological side say that network position is determined by adherence to social norms, which dictates attitudes (the ties that bind and gag), and that the strength of a relationship between the individual and the friend who’s already adopted (their attitude to the friend, if you will) are involved. The latter is a more nuanced, psychological approach, and one which I adopt in my research. Here’s a schematic:

So I’m trying to merge them, to theorise how the psychological influence interacts with the network influence. Here’s what I’ve come up with:

Briefly, what I’m saying is that when an individual is exposed to an innovation, two pathways are activated: social norms and the attitude to the friend. This is the dichotomy between personal and social identity described by the Social Identity Theorists (e.g., Tajfel & Truner), and which some researchers (e.g., Spears & Lea) think is problematic online; they say social and personal identity aren’t possible in online communities. I argue that they are. But that’s another post.

So taking the first pathway, I propose that social norms contribute to both the attitude to the innovation and to an actor’s network position. The reason for the former is quite clear: the social environment influences what is considered a correct or incorrect attitude to have about a new thing. According to Social Identity Theory, how much a person identifies with the group will have an impact on how much s/he adheres to the norms. But that’s another conversation. The reason for the latter draws on Rogers’ theory that people in the centre of networks (i.e., have the most connections) will be less innovative because they must adhere to the social norms of the group if they wish to stay in this important position. If they adopt an innovation that’s outside the norms of the group, they’re in danger of losing friends because they’re too innovative. That’s why there’s a connection between network position and attitude to the innovation.

With reference to the second pathway, I argue that exposure isn’t enough to contribute directly to behaviour; the intention (and ability to adopt the innovation) must be there too (thus the dotted line). However, I also think there’s a stronger link between exposure and behaviour if you take into account the attitude a target has towards the friend who’s adopted: classic psychological studies emphasise the role of interpersonal trust, for example, on influence.

So I’m suggesting here that attitude is a cornerstone towards the adoption of a behaviour, that it feeds the threshold of an individual (i.e., his or her personal tipping point to adopt an innovation – in sturcturalist networks studies this is often only due to exposure), which feeds intention (keeping in line with the TPB) which ultimately contributes to behaviour.

What this theory is really only relevant to is an innovation in which the person is free to choose whether or not s/he adopts it: conformity and compliance aren’t considered here.

That’s the theory. What do you think?


~ by aleks on April 9, 2009.

11 Responses to “On the path towards a merged psychological-network theory of innovation diffusion through online communities”

  1. Hi Aleks – wow…looks like that’s my Easter weekend done trying to get my head round it. It might take me a few weeks; but will endeavour to digest it.

  2. @Niall – you should see the whole thesis 🙂

  3. Hi Aleks

    I don’t know much about psychology and stuff like that, but reading through your post it seemed to make sense to me and struck parallels (for me) of how things happened in World of Warcraft while i was still playing.

    Where new addons were written or third party software used in organising of raids etc and how these were adopted amongst different groups and communities, which sounds similar to what your study is about.

    Definitely think your chart combines the other theories coherently for the context of the analysis, although i did wonder why ‘behavioural efficacy’ was stuck out on its own but then i re-read and it made sense.


  4. Note; I am not a psychologist, but I’ve seen one on TV. Also, my wife has a degree in it, and has therefore been accused of being capable of mind control. But she’s telling me that she doesn’t know I’m writing this, so that’s all right.

    Anyhoo…if I grasp this correctly, you’re claiming that someone’s likelihood of adopting something new is positively correlated with how cool they think those who have already adopted it are. Further, the more coolness attached to someone in a group, the more reluctance they have to adopt things that might be regarded as non-cool by that group. And both of these factors dominate over how cool some innovation can be regarded as being, generally.

    If that’s a proper interpretation, I can see nothing whatsoever to quibble about; if it isn’t, I won’t be the least bit surprised.

    Also, you report this edited claim: “[Spears & Lea] say social and personal identity aren’t possible in online communities.”

    Surely, this generalizes to: “constructed or proxy identities don’t work.”

    If so, Spears & Lea should have no trouble disproving the career of Max Clifford.

  5. Aleks – I’m not saying I could repeat this in the pub 😉 – but I followed it through to the end and it made plenty of sense. That, believe me, is a serious achievement. Nice work.

  6. Hi, Aleks — I’m hopelessly unfamiliar with virtually all the writings you’re citing, but I’d love to read the full thesis when you’re finished.

    One thought on Rogers’ theory. If this is fairly provable empirically, I shouldn’t quibble, but it seems to me that the opposite might be just as true. From an economic-ish perspective, people in the center of social networks could plausibly have considerable social capital to spend on innovation, and thus be more likely to innovate than people at the margins, where a relative paucity of connections makes each one more valuable.

    The relative strength of connection seems like it would also play a role, perhaps more than absolute number. A person with strong social connections would seem more likely to innovate or accept innovations than someone with weak connections.

    Then (sort of free-associating here) — I would imagine there’s some variable that associated with the character or content of the network’s connections. Ie, is the relationship defined by a shared interest in horses, golfing, a long-term deep friendship that transcends individual interests, etc. If innovation itself is a shared value (and status in the community is in some sense associated with innovation or innovation adoption, as it often is in online communities), then maybe network position means less than in culturally conservative communities. Tho this is probably just restating your point about correct or incorrect attitudes.

    Interesting stuff!

  7. I’m thinking that if there is a problem… it is in the different ways the internalist accounts you are using such as ‘intention’ and the externalist accounts ‘network theory’, ‘tipping point’ account for what’s real and/or knowable. I think people turn to network analysis, either actor network or social network, because they can’t give internalist descriptions of events, because they cannot know the state of the mind. They can only know what is reported by the person in the network (though they should not believe the reports), the actions of the persons that they witness, or the actions that are witnessed by others and somehow otherwise verified, etc. I think you see what i’m pointing toward there…. All data is external and mental states are unknown, though possibly reported, can’t be confirmed. So network and social identity theory would be externalist models, we can watch a persons actions and see whether they conform to group identity. I don’t see how you can make tpb work there though because you need to assume something there that is not externalist, you need intent, you need mental states… That i find worrysome because you have to justify the existence of such things in your new model of the subject that you use in your network theory, and i think the justification will likely conflict with the justification for choosing social identity theory and network theory. I’m sure you can work around the problems of introducing internalist objects into externalist analysis, but…. perhaps on the other hand, it might be better to stick with one or the other.

  8. interesting points for me also. I wouldnt quite say that i turned to ANT because i cant know internal fields, but because it allows me to interrogate the external events more fully. My own q is about change and specifically about the uptake of text or sms messaging for counselling. What is the relationship between what people think about change and what they do to enact it. Would love to enter into more discussions on this,

  9. […] next post will extend the theoretical model I posted the other day; upon empirical reflection, it doesn’t quite work in this context. But I have another one – […]

  10. John –
    yes the central person could adopt, and that would result in a very very quick diffusion. Funnily enough, this is what’s happened in this study (although that effect peters out leaving a phase 2 wherein the the diffusion is more s-shaped, suggesting a peripherally-inspired network effect).

    But this is arguably relative to the normative content of the innovation. If it was something like, say, adopting a challenging sexual practice in a conservative group, it’s expected that the central actor would wait until the innovation had been transformed into something that was more acceptable to the social norms.

    Peripheral players are really important too; they may have a paucity of connections with the focal group, but they may also have plenty of connections with other groups.

    Agreed that the number of connections is a short-sighted short-cut, and I’m trying to understand how the (psychological) strength of the relationship might effect the spread of the innovation. Strength is a relative concept, based on the criteria you’re studying; some conceptualise it as multiple interests, like you propose, others think it has to do with the amount of communication that occurs between people. So many factors, so little time!

    @jeremy & alisa – that is, admittedly, what psychologists do. Interesting you think SIT is an externalit perspective; Tajfel & Turner sought to explain when (internally?) an individual would identify with a personal or a social identity. I think I consider this in my next iteration of this model, coming up in less than 5 mins…

    really really great comments folks. Many thanks!

  11. […] path towards a theoretical model: reflections on what came before Last week I posted about a theoretical model I had developed based on three theories – the Theory of Planned Behaviour, the Technology […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: