People are persuaded along two different roads: weighing your actual case, or leaning on cues like who is talking and how it looks. Which road they take depends mostly on how much they care.
The dominant account is the Elaboration Likelihood Model (Petty & Cacioppo, 1986), with a close cousin in Chaiken's Heuristic-Systematic Model (1980). Both say the same core thing in slightly different language: persuasion can happen through careful thought or through shortcuts. The careful path is the central route (the ELM's term) or systematic processing (the HSM's): the person actually evaluates the substance of your case. The shortcut path is the peripheral route, or heuristic processing: the person leans on simple cues, an expert source, an attractive face, a long list of reasons, the sense that everyone agrees, without weighing the argument itself. "Elaboration" is just the jargon for how much mental work the person puts into thinking about the message.
What decides which route runs is the part to remember. The biggest factor is involvement: how personally relevant and important the topic is. The signature experiment (Petty, Cacioppo & Schumann, 1983) showed people a magazine ad for a razor that varied two things at once, the strength of the arguments and whether the endorser was a sports celebrity or an ordinary person, under conditions where the product mattered to them or did not. When involvement was high, argument quality moved attitudes and the celebrity barely registered. When involvement was low, the celebrity moved attitudes and the argument barely mattered. The sister study (Petty, Cacioppo & Goldman, 1981) found the same crossover with source expertise instead of fame. Beyond the situation, people also differ by disposition: those high in need for cognition, the tendency to actually enjoy effortful thinking, take the central route more readily (Cacioppo & Petty, 1982), while distraction or time pressure pushes anyone toward the peripheral one.
Two consequences follow. Peripheral cues genuinely work: a credible source shifts attitudes on its own (Hovland & Weiss, 1951), as do attractiveness, apparent expertise, and sheer number of arguments. But the two routes do not build the same thing. Attitudes formed on the central route, by chewing on real arguments, are more durable, more resistant to later counter-messages, and better at predicting what the person actually does. Attitudes formed on the peripheral route are easier to win and easier to lose.
There is a darker reading of the central route, and it is the one that explains a polarised public debate. Caring enough to think hard does not mean thinking fairly. Petty and Cacioppo themselves separated objective elaboration, weighing a case on its merits, from biased elaboration, thinking hard in order to defend the view you already hold. When a topic is bound up with someone's identity or values, the second kind takes over. People pick apart the arguments they dislike to find the flaw and wave through the ones they like. Lord, Ross and Lepper (1979) showed it cleanly: they handed supporters and opponents of the death penalty the same mixed bag of evidence, and rather than converging, both sides came away more sure they had been right all along. Taber and Lodge (2006) found the same in political argument, with an uncomfortable twist. The people who polarised the most were the most knowledgeable and sophisticated, the ones best equipped to elaborate. More mental horsepower, pointed at refutation instead of truth.
This is why a stronger argument can fail on a heated political issue when the very same argument would win a high-stakes purchase. The buyer is motivated to get it right; the partisan is motivated to stay right. One famous version of this, the "backfire effect" (Nyhan & Reifler, 2010), held that correcting a false political belief could actually strengthen it. That strong claim has not aged well: when Wood and Porter (2019) tested it across 52 issues and more than ten thousand people, corrections usually worked and real backfire was rare. So the honest, better-supported takeaway is narrower than "facts entrench people." On identity-laden topics, a better argument tends to get counter-argued harder rather than accepted, so argument quality quietly stops doing the work it does for the same person when they are simply trying to be accurate.
The sharpest open question is whether there are really two routes at all. The "unimodel" critique (Kruglanski and colleagues, 1999) argues for a single persuasion process in which arguments and cues are both just evidence, weighed according to how relevant they are and how much effort the person is willing to spend, so the tidy two-route split may describe outcomes rather than name two real mechanisms. The ELM also draws fire for being hard to falsify: because the same factor can be slotted into either route after the results are in, the model can explain almost any finding. And that flexibility is built in, because the same variable, say an attractive spokesperson, can serve as a peripheral cue, as an actual argument ("this product is for attractive people like them"), or as something that changes how hard the person thinks, depending on the setting.
The evidence leans heavily on lab studies with students, mock advertisements, and one-off essays, so the route a real campaign travels in the wild is harder to call. And "involvement" is itself a slippery, many-sided idea, which makes the model easier to apply after the fact than to predict with.
Whether persuasion is truly one process or two, how to predict in advance which role a given cue will play, and exactly why central-route attitudes hold up better, are all still live.
The single most useful move here is to read how much your audience cares before you decide how to persuade them, because that sets which road is even open.
Match the message to involvement. For a high-stakes, considered purchase (a car, enterprise software, anything expensive or identity-laden), buyers scrutinise, so you win on the central route: real arguments, evidence, substance. A glossy celebrity spot will not survive that scrutiny. For a low-stakes or habitual purchase (a snack, a cheap impulse buy), people are not weighing much, so peripheral cues do the work: attractive design, a likeable face, social proof, ease, familiarity. The expensive mistake is mismatching, spending on a celebrity to sell scrutinised software, or burying an impulse product in a dense spec sheet. And because central-route belief lasts while peripheral belief fades, anything you want people to hold onto needs something real to think about.
On most issues most voters are low-involvement, so peripheral cues carry the message: the messenger, identity, tone, repetition. For the engaged base and genuinely high-stakes issues, arguments do more work. Match the message to the segment's involvement rather than broadcasting one register to everyone. There is also an identity trap to respect: on issues that have become markers of who someone is, piling on facts and sharper arguments tends to harden the people who already disagree, because they elaborate to defend rather than to learn. Movement there comes less from better evidence than from a trusted messenger, shared ground, and dialling down the sense of threat to identity.
For decisions you want people to hold to under pressure (vaccination, pensions, long-term health behaviour), invest in central-route persuasion: clear arguments and the room to think them through, because that belief resists later counter-messaging. Peripheral nudges can move the moment, but they fade, so they are better for one-off actions than for durable change. When you need to correct a false belief, the encouraging news is that simply giving people the correction usually does move them (Wood & Porter, 2019); the stubborn cases are beliefs fused with identity, where who delivers the correction matters more than how strongly it is argued.
Before choosing a message, judge two things about your audience: how much they care, and whether they have the capacity (time, attention, expertise) to think it through. High on both, persuade on the merits and it will last. Low on either, lean on cues and accept that the effect is shallow and temporary. And when you need belief to stick, give people a real argument to chew on, not just a reason to nod.
.png)