Resources
Social science and human behaviour: why people adopt, resist, trust, and abandon technologies, and why none of it shows up in the specification, each idea tied to the literature and, where we have one, a case.
The Lab's other theory pages take the view from above. The economics of transitions asks who bears a cost and who captures a value. The technology & innovation dynamics page asks how a technology evolves and how firms compete to own it. The transitions primer asks how whole systems move. This page takes the view from the ground: what happens when a real person, in a real place, actually meets the technology.
It is the lens the Lab leads with. Social science first, technical depth alongside, is not a slogan; it is a claim about where transitions are actually decided. A payment app can be cheaper, faster, and better in every measurable way and still fail because people do not trust it, cannot fit it into how they already live, or were never the people the designers pictured. The reasons are behavioural, social, and cultural, and they are invisible to a specification, a cost model, or a strategy deck. They become visible only by going and asking.
Each concept is defined, anchored to its canonical source, and, where the Lab has written about it, linked to the applied case. Full references are at the end. Running underneath all of it is one methodological point: because these forces cannot be read from a distance, the Lab reads them in the field.
People do not optimise; they satisfice. Faced with limited time, information, and attention, they reach for a decision that is good enough rather than the best one, and they use rules of thumb to get there. Simon (1955) named bounded rationality, and it is the foundation of everything in this cluster: a transition designed for the rational calculator of the textbook will misjudge the actual human being, who is busy, uncertain, and working with far less information than the model assumes.
People feel losses roughly twice as heavily as equivalent gains, and they judge outcomes against a reference point rather than in absolute terms. Kahneman and Tversky (1979) formalised this as prospect theory, and it reshapes how a transition should be framed: asking a household to give up something certain (the familiar cooker, the petrol bike) for a larger but uncertain future gain runs straight into loss aversion. The switch is not resisted because people cannot do the maths; it is resisted because the maths is not how people weigh a loss. This is the behavioural cousin of the economic argument on the transition trough.
The same choice, described two different ways, produces two different decisions, and people file money and costs into separate mental accounts that they treat differently. Tversky and Kahneman (1981) showed this for framing. A daily payment and an equivalent annual one feel different; a saving framed as avoided loss lands harder than one framed as a gain. For a transition this means the framing of an offer, and how it fits a person's mental accounts, can matter as much as its actual terms, which is why the Lab's interviews probe how people describe a cost, not just what it is.
People systematically overweight the present and underweight the future, and they do so inconsistently, being patient about the distant future but impatient right now. This is why a technology with high upfront cost and delayed benefit, efficiency, solar, insulation, is under-adopted even when the long-run case is overwhelming: the upfront cost is felt now, the benefit is discounted away. Pay-as-you-go financing works precisely because it moves the cost into the same present as the benefit.
People struggle to value a saving relative to a future that will not happen, because the avoided cost is invisible, it never arrives to be noticed. This behavioural fact sits underneath a market failure the Lab has written about: efficiency measures with fast paybacks go unbought because the benefit is a non-event. Set out in Nobody Buys a Chiller.
Whether someone adopts a technology turns on two perceptions above all: how useful they think it is, and how easy they think it is to use. Davis (1989) built the technology acceptance model around perceived usefulness and perceived ease of use, and its enduring lesson is that adoption is driven by perception, not by the objective properties of the device. A technology that is genuinely useful but perceived as hard, or genuinely simple but perceived as pointless, will not be adopted. What people believe about a technology is the thing to measure.
Later work widened the picture. Venkatesh and colleagues unified the acceptance literature into four drivers: performance expectancy (will it help), effort expectancy (is it hard), social influence (do people I respect use it), and facilitating conditions (do I have the support and infrastructure to use it). Venkatesh et al. (2003) drew the picture together. The third and fourth are the ones engineers most often forget: adoption is social and infrastructural, not just individual, and a good product with no social proof or no support around it stalls.
Intentions, not attitudes, predict behaviour, and intentions are shaped by three things: a person's own attitude, the social norm they perceive, and how much control they feel they have. Fishbein and Ajzen (1975) and Ajzen (1991) built the theories of reasoned action (TRA) and planned behaviour (TPB) on this, and the practical consequence is sharp: changing what people think of a technology is not enough if the perceived norm is against it or if people feel they cannot actually act. Perceived behavioural control, whether a person believes they can do the thing, is often the missing piece.
An innovation spreads through a population in a sequence, innovators, early adopters, early and late majority, laggards, and each group adopts for different reasons and responds to different evidence. Rogers (1962) mapped this. The behavioural lesson, and a recurring field finding, is that the enthusiast who adopts first is not representative of the majority who must adopt later, so evidence gathered from early users systematically misleads about how the mainstream will respond. The competitive-dynamics reading of the same curve is on the technology & innovation dynamics page.
People rarely replace one thing cleanly with another. They add the new alongside the old, keep the incumbent as a fallback, and shift gradually as trust builds, so real adoption looks like stacking, not switching. We read this behaviour in African solar in Stacking, Not Switching.
Before a person will use a technology that touches their money, their health, or their safety, they must trust it, the device, the provider, and the system behind it, and trust is slow to build and fast to lose. Much of what looks like irrational resistance is a rational withholding of trust from a system that has not yet earned it. In many of the Lab's field settings, trust, not price or performance, is the binding constraint on adoption, and it is built through experience and social proof, not through specifications.
People do not assess risk the way an actuary does. Perceived risk is shaped by whether a hazard is familiar, voluntary, controllable, and fairly distributed, so a small but dreaded or imposed risk can weigh more heavily than a large but familiar one. Slovic (1987) mapped these dimensions. A transition that dismisses public concern as innumeracy misreads it: the concern is usually about control, fairness, and unfamiliarity, and it has to be engaged on those terms, not corrected with a statistic.
A technology needs not only to work but to be seen as legitimate, by users, communities, and authorities, or it will be resisted regardless of its merits. Legitimacy is a social judgement about whether a technology and its provider have the right to do what they are doing, and it can be withheld even from something that performs well. In extractive-industry practice the same idea is named the social licence to operate. Reading who confers legitimacy in a given setting, and on what terms, is often the difference between a pilot that is tolerated and one that is embraced.
A system that works well on average can fail badly for a specific group, and when it does, it forfeits that group's trust entirely. A confident aggregate accuracy figure can conceal exactly the failures that matter most to the people a technology was meant to include. Set out for AI systems in Who Does It Fail For?.
A technology's design is not dictated by the technology itself; different social groups read the same artefact differently, and the final form reflects whose interpretation won, not what was technically inevitable. Pinch and Bijker (1984) called this the social construction of technology (SCOT), and it dissolves the idea that a design is simply the best solution: it is the solution that suited the groups with the power to shape it. For a transition, this means the "obvious" form of a technology is a social outcome that could have been otherwise, and can still be contested. This is the ground-level version of the social shaping the transitions primer treats at system scale.
The same technology means different things to different people, and those meanings, not the object alone, drive how it is used, resisted, or repurposed. A mobile phone is a status object, a livelihood tool, a safety device, or a distraction depending on who holds it. A transition that assumes everyone reads the technology the way its designers do will be surprised by how it is actually taken up, which is why the Lab asks what a technology means to people, not just whether they use it.
People do not passively accept technologies as given; they modify, misuse, and repurpose them to fit their own lives, and this appropriation is a form of design in its own right. The workaround, the unofficial use, the hack, these are signals of what people actually needed, which the original design missed. Watching how a technology is appropriated in the field is one of the richest sources of insight the Lab has, because it reveals the gap between what was built and what was wanted.
Big technological systems, grids, road networks, payment rails, accumulate momentum: as they grow, the sunk investment, the institutions, and the habits built around them make their direction progressively harder to change. Hughes (1983) showed this for electrification, and the frame is known in the literature as large technical systems, or LTS. It is the human-and-institutional counterpart of lock-in: the system resists redirection not only for economic reasons but because so many people's routines, skills, and expectations are bound up in it.
Much of what a technology has to change is not a decision but a practice, a routine bundle of materials, competences, and meanings that people perform without deliberating: how a household cooks, travels, washes, keeps warm. Shove and colleagues (2012) argue that these practices, not individual choices, are the real unit of change, and that a transition succeeds only when the whole practice is reconfigured, not when a single product is swapped in. This is why giving someone a cleaner cooker often fails: the cooking practice, the fuel, the pot, the timing, the taste, the skill, did not change with it.
Most everyday technology use is habitual, performed automatically with little conscious thought, which is why information campaigns so often fail to change behaviour: they target the deliberating mind, but the behaviour lives in habit. Behaviour change usually requires disrupting the routine or changing the context, not simply supplying better facts. A transition strategy built on "if people only knew" tends to underperform, because knowing is not where the behaviour comes from.
Because so much behaviour is automatic, the way choices are arranged, what is the default, what is easiest, what is most visible, shapes what people do, often more than persuasion does. Thaler and Sunstein (2008) called this choice architecture, and its most powerful tool is the default: people tend to stick with whatever they are given unless they actively opt out. For a transition, setting the sustainable option as the default, rather than exhorting people to choose it, is frequently the higher-leverage move.
When a new technology asks for more effort than the incumbent, even a small amount more, it tends to lose, because convenience is weighed far more heavily than people admit or than merit would justify. Many transitions fail not on cost or performance but on friction: an extra step, a wait, a hassle. The Lab's fix-it question in the field, what would you change first, so often surfaces a friction, not a feature, because friction is what actually governs everyday use.
Designers build for a user they picture, and that imagined user is usually more literate, more connected, more moneyed, and more like the designer than the real population. The gap between the imagined user and the real one is where exclusion is manufactured, quietly, at the design stage, long before deployment. The Lab's core method, going to the people a technology actually reaches, exists precisely to close this gap, and its second-perspective work exists because the people who collect feedback from their own users rarely find the ones they never reached.
A technology can exclude not through malice but through assumptions baked into it: an interface that assumes literacy, a service that assumes a bank account, a device that assumes reliable power. These assumptions are invisible to those who meet them and decisive for those who do not. Reading a technology for its embedded assumptions, and asking who they quietly shut out, is a standing part of the Lab's fieldwork.
People tend to tell a researcher, especially one associated with the provider, what they think the researcher wants to hear, so feedback collected by an organisation from its own users runs systematically positive. This is a behavioural fact about the interview itself, and it is the reason independent data collection produces materially different results from self-report. It is the human-behaviour foundation of the economic argument in The Reporting Loop.
A technology adopted by a household is not adopted equally by everyone in it. Who controls it, who benefits, and who bears the burden often divide along lines of gender, age, and power inside the home, so a headline adoption figure can hide a very unequal reality. A clean cooker bought by a household may change one member's day and not another's. Reading impact at the level of the person, not the household, is how these distributions become visible.
The concepts above draw on the following canonical sources in behavioural economics, social psychology, technology acceptance, and the sociology and anthropology of technology.
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179 to 211.
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319 to 340.
Fishbein, M., and Ajzen, I. (1975). Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research. Reading, MA: Addison-Wesley.
Hughes, T. P. (1983). Networks of Power: Electrification in Western Society, 1880 to 1930. Baltimore: Johns Hopkins University Press.
Kahneman, D., and Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263 to 291.
Pinch, T. J., and Bijker, W. E. (1984). The social construction of facts and artefacts. Social Studies of Science, 14(3), 399 to 441.
Rogers, E. M. (1962). Diffusion of Innovations. New York: Free Press.
Shove, E., Pantzar, M., and Watson, M. (2012). The Dynamics of Social Practice: Everyday Life and How It Changes. London: SAGE.
Simon, H. A. (1955). A behavioral model of rational choice. Quarterly Journal of Economics, 69(1), 99 to 118.
Slovic, P. (1987). Perception of risk. Science, 236(4799), 280 to 285.
Thaler, R. H., and Sunstein, C. R. (2008). Nudge: Improving Decisions About Health, Wealth, and Happiness. New Haven: Yale University Press.
Tversky, A., and Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453 to 458.
Venkatesh, V., Morris, M. G., Davis, G. B., and Davis, F. D. (2003). User acceptance of information technology: Toward a unified view. MIS Quarterly, 27(3), 425 to 478.
This is the fourth of the Lab's theory references, and in practice it is the one that comes first, because it is the lens the Lab leads with.
The four pages overlap on purpose. Diffusion of innovations appears here as human behaviour and on the innovation page as competitive dynamics; loss aversion here has a cousin in the economics of the transition trough; the social construction of technology is the ground-level version of the social shaping the transitions literature studies at the system scale. Each page treats the shared idea through its own lens.
The reason this page comes first is methodological. Everything on it, perception, trust, habit, meaning, appropriation, exclusion, has one property in common: it cannot be read from a distance. It does not appear in a specification, a cost model, a patent, or a dashboard. It appears only when you go to where the technology meets the person and ask, carefully and independently, what actually happened. That is what the Lab does, and this page is the theory of why it is necessary. For how we put it into practice, see What We Do and the In-Depth Interview Guide.