{"id":2955,"date":"2023-12-24T15:11:42","date_gmt":"2023-12-24T15:11:42","guid":{"rendered":"https:\/\/illuminated-text.com\/cb\/?p=2955"},"modified":"2025-08-14T13:11:13","modified_gmt":"2025-08-14T13:11:13","slug":"involvement-persuasion-elm","status":"publish","type":"post","link":"https:\/\/illuminated-text.com\/cb\/involvement-persuasion-elm\/","title":{"rendered":"Involvement & Persuasion"},"content":{"rendered":"\n

(header image credit<\/a>)<\/p>\n\n\n\n

A quick introductory puzzle?<\/h4>\n\n\n\n

Say you write a better textbook. It explains things better than its competitors, is neurodiversity-friendly, etc#up.<\/p>\n\n\n

\n
#down. Now who would do a thing like that...<\/pre>\n<\/div><\/div>\n\n\n
In a logical world, professors would assign it to their classes as soon as they heard of it.<\/p>\n\n\n\n
In the real world? Not so easy. People choose worse products over better ones. You do it all the time, most likely. So do I.<\/p>\n\n\n\n
Why does the best thing often lose? Why is persuasion so hard?<\/p>\n\n\n\n
A recipe for guaranteed persuasion?<\/h2>\n\n\n\n
Anyone with a creative mind (and\/or a chatbot) can write a persuasive message selling a textbook (“education is important, let’s give our students our best!”). But which of those messages is best?<\/p>\n\n\n\n
If there were one correct answer, a guaranteed method, we would see it every day of our lives from every single advertiser. But we don’t, because there isn’t.<\/p>\n\n\n\n
As with nearly every worthwhile question about humans, the real answer is: “It depends”, and the real question is: “Ok smartypants, what does it depend on?”<\/p>\n\n\n\n
Bad news: The answer is “many things”.<\/p>\n\n\n\n
Good news: There are a few things that matter often and a lot, and we can know about them. Sure they don’t matter every time, but when it comes to predicting humans, “often” is pretty good#up. <\/p>\n\n\n
\n
#down. As George Box famously put it:<\/a> \"all models are wrong, but some are useful\".<\/pre>\n<\/div><\/div>\n\n\n
One of the most consistent factors is involvement<\/strong>: how much does the audience care about your message at the moment they hear it?<\/p>\n\n\n\n
Involvement and persuasion… it matters?<\/h3>\n\n\n\n
More and more.<\/p>\n\n\n\n
In the 1950’s you could stick an ad full of words into a magazine, and people would read it:<\/p>\n\n\n
\n
\"1949
Remember this ad from chapter 4? Side note: In the 1950’s you could also make wildly unsupported medical claims and no regulators would demand proof (credit<\/a>)<\/figcaption><\/figure>\n<\/div>\n\n\n
Advertising wasn’t new. They had seen it before#up (“how beguiling darling, this is the very soap from the market!”). There was just a lot less of it back then. The cleverest and most talented people of their age had not yet built careers on using ads to monetize every possible part of our lives (“today’s meeting of Citizens Against Advertising is sponsored by…”).<\/p>\n\n\n
\n
#down. Advertising has been around since at least the 1800s<\/pre>\n<\/div><\/div>\n\n\n
You probably see hundreds of adverts every day. We bet you are pretty good at ignoring them, too. You couldn’t survive without blocking them out. The lesson: Most ads are half-watched at best. Mostly they exist as background noise in our lives.<\/p>\n\n\n\n
But does that mean that advertising doesn’t work? That we throw it into an uncaring world, where it is immediately forgotten?<\/p>\n\n\n\n
If advertising didn’t support sales, then companies would have stopped paying for it. Yet around the world, they spend hundreds of billions of US dollars on it every year.<\/p>\n\n\n\n
But why? Does it matter how much attention the audience pays? Can we get people to pay more attention?<\/p>\n\n\n\n
The Elaboration Likelihood Model<\/h2>\n\n\n\n
We are going to start with a partial wrong turn.<\/p>\n\n\n\n
That’s not something teachers normally do#up, but some theories are so big and famous that you have to tackle them anyway. Our most famous persuasion theory is the Elaboration Likelihood Model, or “ELM” <\/span>, and it says there are two key ingredients <\/strong>for how well a persuasive message will work:<\/p>\n\n\n
\n
#down. Not on purpose<\/pre>\n<\/div><\/div>\n\n\n
    \n
  1. The level of involvement <\/strong>that people have with it<\/li>\n\n\n\n
  2. Whether the message makes a strong or a weak argument<\/strong>#up, or whether it just uses persuasive cues<\/strong><\/li>\n<\/ol>\n\n\n\n
    The first of those ingredients is almost certainly right. The second one, it turns out, probably isn’t. Or, at least, not in quite the way that the ELM described it. We’ll talk about how to fix it in a minute.<\/p>\n\n\n
    \n
    #down. In normal life an 'argument' means shouting at each other. But we use the academic meaning, which is when you give reasons to believe a thing.<\/pre>\n<\/div><\/div>\n\n\n
    How the ELM’s says it works<\/h3>\n\n\n\n
    It assumes that you try to form accurate beliefs when someone tries to persuade you, but<\/strong> that you can be paying more or less attention while doing this. Here’s what it looks like<\/p>\n\n\n
    \n
    \"A
    The ELM<\/figcaption><\/figure>\n<\/div>\n\n\n
    When involvement is HIGH, according to the ELM<\/h5>\n\n\n\n
    The ELM claims that we use “central route” processing when we are sufficiently motivated and given enough time and space. This form of thinking has us behaving like good academics: searching for arguments to interrogate, and discarding cues#up as unconvincing.<\/p>\n\n\n
    \n
    #down. We don't explain arguments vs. cues here as they turn out not to be important. But an accordion box which explains them in a few more paragraphs if you want to know<\/pre>\n<\/div><\/div>\n\n\n
    If our central processing finds an argument that stands up to scrutiny, with solid facts woven into logical conclusions, then we are persuaded. Our attitude is changed. Even better, this extensive mental elaboration should help us encode those changed beliefs into our long-term memory, creating persuasion that lasts over time. <\/p>\n\n\n\n
    If, on the other hand, we decide the arguments are weak then we reject them, and sometimes even reject the entire persuasion attempt itself (“Well that’s just dumb!”), meaning that it backfires, and we encode negative thoughts about it into our memory.<\/p>\n\n\n\n
    When involvement is LOW, according to the ELM<\/strong><\/h5>\n\n\n\n
    We engage in “peripheral route” reasoning. The claim is that peripheral processing isn’t sophisticated enough to pay attention to arguments, but may still notice simple prompts called ‘cues’. The ELM assumes that these are easy to understand quickly and effortlessly.<\/p>\n\n\n\n
    The ELM claims that peripheral processing is not deep enough to encode new meanings into our long-term memory. It may win us over, but only for a few minutes and then we forget all about it.<\/p>\n\n\n\n
    That all sounds reasonable. What’s wrong with it?<\/h3>\n\n\n\n
    Like any good scientific theory, the ELM has evolved as new evidence has emerged <\/span>. Unlike a good theory, the ELM has a deeper problem: Some of its key predictions turn out to be very unreliable. It turns out that Low-involvement people don’t stick to being strictly influenced by cues, and high-involvement people don’t stick to only processing strong arguments <\/span>.<\/p>\n\n\n\n
    In science-speak, the ELM doesn’t “replicate” well. People who try to make it happen again tend to come up empty: This database<\/a> tracks 20 such recreation attempts#up (as of December 2024) and all of them are found to be “informative failures to replicate” – meaning that they found interesting things, but not the ELM’s predictions for what should happen.<\/p>\n\n\n
    \n
    #down. Click \"Replication tracker\" at the very top of the page, and then search for \"Petty\" (Rich Petty is one of the original study authors).<\/pre>\n<\/div><\/div>\n\n\n
    The ELM thus appears to be a victim of academia’s widespread “replication crisis” (as described in<\/a> our intro chapter).<\/p>\n\n\n\n
    “Yes, fine” you say, “but couldn’t you have skipped telling me all that wrong stuff?”<\/p>\n\n\n\n
    That’s fair. But here are a few reasons:<\/p>\n\n\n\n
    1:<\/strong> The ELM is so famous that you are going to hear people talking about it. You need to know what is going on.<\/p>\n\n\n\n
    2:<\/strong> Involvement does seem to matter, just not the way the ELM thought it did. We can build on that.<\/p>\n\n\n\n
    3:<\/strong> It’s a great example of why ideas get popular: It makes intuitive and practical-sounding predictions, with just the right amount of complexity to hit a psychological sweet spot: Not too simple that it feels obvious, but still simple enough that it can be learned quickly. Watch out for this in future, and set your scepticism levels appropriately.<\/p>\n\n\n\n
    Ok, but I still want to know more about what ‘arguments’ and cues are supposed to be<\/h4>
    \n
    In normal English “arguments” means fighting with someone, but in Academia it means trying to persuade by giving reasons. Cues are what they sound like – prompts that gently direct you in the desired direction. <\/p>\n\n\n\n
    To sell cupcakes with a cue you might say: “they are made with all natural ingredients” (hey, they sound healthy!). To sell them with an argument, you might say “made with psyllium fibre, which has been shown to reduce the risk of heart attacks” (hey, that’s a good reason!).<\/p>\n\n\n\n
    \n
    \n
    Argument<\/strong><\/p>\n<\/blockquote>\n\n\n\n
    An attempt to persuade people by providing some kind of reason or justification<\/p>\n<\/div>\n\n\n\n
    \n
    \n
    Cue<\/strong><\/p>\n<\/blockquote>\n\n\n\n
    Prompts that try to lead people to do or believe something<\/p>\n<\/div>\n\n\n\n
    Arguments – give me more examples!<\/h4>\n\n\n\n
    Recall that in academia an ‘argument’ is only about reasons and justifications. If your roommate shouts at you that you never take out the recycling, that is not an argument, just fighting. But if they say: “it’s your turn because I took it out last week”, then that is<\/strong> an argument because it has 1) persuasive intent, and 2) gives some kind of justification (whether you believe it or not). Those are the two key ingredients.<\/p>\n\n\n
    \n
    \"Poster
    A simple argument: Buy our shoes because they will keep your feet dry (credit<\/a>#up)<\/figcaption><\/figure>\n<\/div>\n\n
    \n
    #down. FYI this image is sourced from an ad agency trying to do content marketing on social media by showing their audience examples of good ads<\/pre>\n<\/div><\/div>\n\n\n
    This shoe ad has those two ingredients: persuasive intent and reasoning. Its reasoning is not stated explicitly: It doesn’t say: “Our shoes are more waterproof than others, and so they will keep your feet drier, and wet feet are no fun”. But it implies this so strongly that the audience sees this meaning anyway. That is often how good advertising makes its arguments.<\/p>\n\n\n\n
    Cues – give me more examples<\/h4>\n\n\n\n
    Cues are a very <\/strong>broad catch-all term for persuasion that doesn’t try to use reasoning. That sounds weird, but you see it all the time.<\/p>\n\n\n\n
    For instance, see if you can spot any arguments in this advert:<\/p>\n\n\n
    \n
    \"A
    Reasons? Justifications? No? (credit<\/a>)<\/figcaption><\/figure>\n<\/div>\n\n\n
    They arranged a tub of this beard product aesthetically with some old-fashioned items. This doesn’t give you a single reason to believe anything, so there is no argument. And yet it is still an attempt to persuade. Possibly even an effective one, for some people.<\/p>\n\n\n\n
    The beard cream ad suggests that their product shares a style conveyed by that scene. It doesn’t say why or what you should do about this. It simply arranges the things together and hopes this inspires you to like or identify with it.<\/p>\n\n\n\n
    Here is another example:<\/p>\n\n\n
    \n
    \"Digital
    A digital poster at a London Underground train station (credit<\/a>)<\/figcaption><\/figure>\n<\/div>\n\n\n
    Again, there isn’t an argument. It doesn’t tell you that COVID vaccines will save your life (or other people’s lives by not passing them a disease). It just suggests that getting the vaccine is normative – it’s what “millions” of people do – and so that makes it more appealing as a thing that you should do too.<\/p>\n<\/div><\/div>\n\n\n\n
    Most textbooks would discuss heuristics and emotions around this point<\/h4>\n\n\n\n
    They are usually described as common types of cues. We will discuss them elsewhere<\/a>, because cues don’t seem the most useful way of thinking about involvement and persuasion.<\/p>\n\n\n\n
    Wait, I’ve heard of a “Heuristic-Systematic model”. This sounds very similar…<\/h4>
    \n
    My goodness, that’s very scholarly of you!<\/p>\n\n\n\n
    The HSM, or Heuristic-Systematic Model <\/span>, is extremely similar to the ELM. They were both developed in the early 1980s, and describe the same two routes to persuasion, although the HSM gives them slightly more accurate names. Instead of “peripheral” it refers to the “heuristic” route because it involves using mental shortcuts, and in place of “central” it talks about the “systematic” route because it involves evaluating arguments systematically (i.e., carefully and thoroughly).<\/p>\n\n\n\n
    The substantive differences between them are small and mostly of historical interest for our purposes. This was a case of two teams describing roughly the same thing, and so coming up with similar descriptions.<\/p>\n\n\n\n
    For an example of their differences, the HLM says that a message can be processed both systematically and heuristically at the same time, but the ELM says you pick one or the other. For what it’s worth, research here has generally favoured the HSM’s take <\/span>.<\/p>\n\n\n\n
    This page could easily reference the HSM instead of the ELM, but we picked the better-known of the two.<\/p>\n\n\n\n
    Why is the ELM better known? There are innocent explanations: Maybe it was promoted better. Maybe it was easier to explain to broad audiences.<\/p>\n\n\n\n
    There is also a darker possibility: That the ELM was created by men (Richard and John), and the HSM by women (Shelley and Alice). Evidence on this is mixed, but some studies do suggest that women can be cited less than men <\/span>, and, less controversially, that women can be held to higher standards of evidence to show competence than men <\/span>. It probably didn’t help that when this research was done (in the 1980’s) there was a stronger popular stereotype of women as incompetent than men than there is today <\/span>.<\/p>\n<\/div><\/div>\n\n\n\n
    But wait… It’s not all wrong?<\/h3>\n\n\n\n
    Kruglanski and Thompson <\/span> suggest a “unimodel”#up that combes arguments and cues together.<\/p>\n\n\n
    \n
    #down. \"uni\" means \"one\" - think of a unicorn with one horn. Even so, the word \"unimodel\" is pretty confusing; This is why you don't let academics name important things<\/pre>\n<\/div><\/div>\n\n\n
    Let’s draw out what this might look like:<\/p>\n\n\n\n
    Original Elaboration Likelihood Model<\/th>A unimodel version<\/th><\/tr>
    \"A<\/td>\"The<\/td><\/tr><\/thead><\/table>
    Table 1: Original and updated versions of the ELM<\/figcaption><\/figure>\n\n\n\n
    It’s mostly the same, except we get rid of “arguments” and “cues” and talk instead about processing quickly or in depth. Also, in place of “attitude change”, we get a mouthful about “change or reinforce beliefs and\/or memory”.<\/p>\n\n\n\n
    Wait a minute… Quick and slow processing… Is this just System 1 and System 2?<\/h4>\n\n\n\n
    Yes!<\/p>\n\n\n\n
    Recall in Chapter One<\/a> we said that low involvement is associated with System 1 thinking and high involvement with System 2? The ELM was one of the very first ‘dual process’ models. Eventually, more and more of them were proposed (“what if the way we judge other people is sometimes quick and effortless, but sometimes more thoughtful…”). Eventually, psychologists merged them all into a general System 1 and 2 that work across many areas of life.<\/p>\n\n\n\n
    Reminder: nobody thinks there is one single part of your brain that does System 1 thinking. These systems are just different ‘modes’ of thought that follow broadly different patterns.<\/p>\n\n\n\n
    What does this look like in reality?<\/h2>\n\n\n\n
    Our agenda for the rest of this page:<\/p>\n\n\n\n
      \n
    1. Persuasion at low levels of involvement<\/li>\n\n\n\n
    2. Persuasion at high levels of involvement<\/li>\n\n\n\n
    3. Changing levels of involvement<\/li>\n<\/ol>\n\n\n\n
      Persuasion at low levels of involvement <\/h3>\n\n\n\n
      How do you advertise to people who are barely listening to you?<\/p>\n\n\n\n
      The Mere Exposure Effect<\/a> may seem like an obvious choice, but it is very weak and limited, as we will learn soon. Its impact on people is small, people have to be exposed dozens of times, and it only works on simple stimuli, like logos – it doesn’t apply to complex stimuli like a truth-claim (such as “my product is safe!”).<\/p>\n\n\n\n
      So let’s say you do have a message. Maybe you sell padlocks, and you want to run adverts telling people that yours are very secure#up. You can run ads telling people this, but is there any point? Most people who see the message won’t be thinking of buying a lock at that point in time, so your explanations will be part of the commercial background noise that fills their lives. <\/p>\n\n\n
      \n
      #down. Security seems like a sensible positioning for a lock company, but see Byron Sharp's take on the virtues of meaningless distinctiveness<\/a>.<\/pre>\n<\/div><\/div>\n\n\n
      We will discuss two reasons that your messages might work anyway, if you do them right: The Truth Effect, and Belief Inferences.<\/p>\n\n\n\n
      The Truth Effect<\/h4>\n\n\n\n
      People absorb claims as true simply from hearing them repeatedly, even without any explanation. As Hawkins et. al. explain:<\/p>\n\n\n\n
      \n
      [a] simple repetition of claims is sufficient to change people’s beliefs…. [A series of] studies [(Arkes, Hackett, & Boehm, 1989; Bacon, 1979; Hasher, Goldstein, & Toppino, 1977)] presented participants with general knowledge statements from various categories (e.g., politics, science,and art) and found that participants rated repeated statements as more valid than non-repeated statements. This finding is known as the truth effect.<\/p>\n<\/span><\/cite><\/blockquote>\n\n\n\n
      \n
      \n
      The Truth Effect<\/strong><\/p>\n<\/blockquote>\n\n\n\n
      When people hear claims repeated over and over they tend to start experiencing them as more plausible and true<\/p>\n<\/div>\n\n\n\n
      This is related to the Ebbinghaus effect#up, which (despite its complicated-sounding name) is one of the oldest and simplest in all of psychology: we can learn things (even nonsense syllables like “spog” and “vlet” and “wab”) simply by seeing or hearing them over and over<\/p>\n\n\n
      \n
      #down. Hermann Ebbinghaus was a German scientist who died in 1909 and was famous for experimenting on himself by reading lists of nonense syllables, and then tracking how quickly he forgot them. His forgetting appeared to accelerate quickly over time unless he reinforced his memory with repeated exposures to those syllables, spaced out over time <\/span><\/pre>\n<\/div><\/div>\n\n
      \n
      \"A<\/a>
      Keep repeating the belief that Coke is refreshing, and perhaps people will feel it to be true (credit<\/a>)<\/figcaption><\/figure>\n<\/div>\n\n\n
      The Truth Effect can be surprisingly powerful. One study showed, for instance, that repeating climate skeptic claims at people can increase their beliefs, even for people who endorse climate science <\/span>. There is an extra-important point in this: Persuasion techniques can be used for ill#up as well as good. On their own they are only tools.<\/p>\n\n\n
      \n
      #down. For a beautifully well-researched and engagingly presented history of the climate skeptical movement (where it comes from and how wildly inaccurate it is) Origin Story has a lovely podcast episode<\/a><\/pre>\n<\/div><\/div>\n\n\n
      A large portion of the advertising you see (and increasingly a lot of politics too) is built on the truth effect. Claims are just repeated over and over until they seem to people like just the way the world is.<\/p>\n\n\n\n
      Belief Inferences<\/h4>\n\n\n\n
      The truth effect might get people to feel that your brand of lock are safe, but can we do even better? Recall that strong attitudes<\/a> are longer lasting than weak ones, resist change better, and guide behaviour more. Can we get that too?<\/p>\n\n\n\n
      Look back at what makes attitudes strong: Part of it is to do with beliefs that fit the purchasing context, and part of it is about building relevant emotions. But a third element is about people having not just one belief, but a complete and suitable set. <\/p>\n\n\n\n
      Perhaps your lock company starts with a simple claim about why its locks are so secure: they are the hardest to pick in the industry. Later, it switches to a different claim that supports the same point: that its keys are very difficult to copy. Later still, it switches to a third supportive belief. Evidence suggests that even low-involvement audiences do seem to eventually put these messages together, and draw the higher-level inference: that your locks are secure <\/span>. This is the belief inference effect.<\/p>\n\n\n\n
      \n
      \n
      The Belief Inference Effect<\/strong><\/p>\n<\/blockquote>\n\n\n\n
      When people hear several simple claims repeated that all support the same belief, they eventually draw the inference that the belief is true<\/p>\n<\/div>\n\n\n\n
      If this logic sounds familiar to you, it should:<\/p>\n\n\n\n
      We talked about a similar idea in chapter one, in which you offer “reason-to-believe” for a product’s core benefit. What we add here is details on how you do this – i.e., slowly and repeatedly drip-feeding people these reasons to believe, until they eventually assemble those reasons into a more abstract belief about the brand’s virtues.<\/p>\n\n\n\n
      Far more recently, we described the Theory of Planned Behaviour, which says (among other things) that people mentally add together their beliefs to form overall attitudes. Again, the logic is very similar: if I believe that your locks are strong, that they are easy to use, and keep things safe then I should form a positive attitude towards your brand#up.<\/p>\n\n\n
      \n
      #down. Assuming I also see those things as important - so says the TPB<\/pre>\n<\/div><\/div>\n\n\n
      A tiny caveat<\/h5>\n\n\n\n
      To be fair, much of the evidence for these effects was gathered in the type of psychology laboratory setting that Byron Sharp criticises as unrealistic <\/span>… But they are also accurate descriptions of much real-world advertising. We often see simple claims repeated over and over, with the apparent belief that they will eventually start to ‘feel right’ to us.<\/p>\n\n\n\n
      High-involvement persuasion: <\/h3>\n\n\n\n
      Consumers may not often pay close attention to commercial messages, but it can be very handy when they do. This is especially true for products whose benefits are hard to describe. Think, for example, of medical products (“this health app maps the glycemic index of foods you eat against your energy needs each day!”), or very technical ones (“certain classes of car may benefit from our new rear differential that reduces mechanical stresses…”). If you can’t break the benefit down into a short, clear phrase, then you might need to find a time when consumers will slow down and pay proper attention to an explanation.<\/p>\n\n\n\n
      High involvement can also be useful for another reason: Again, recall<\/a> the benefits of strong attitudes. Part Building up that complete and suitable set of beliefs, and tying it to the right emotions, can be easier when you have more of people’s attention. If they are willing to invest more mental effort, they will more easily form dense networks of ideas about your product, linking its attributes to supportive ideas, beliefs, and contexts. Any proper enthusiast will know far more about car (or bike or knitting) brands than you will ever fit into a handful of half-ignored 30-second advertising spots.<\/p>\n\n\n\n
      So… what would a high-involvement ad look like?<\/h5>\n\n\n\n
      Maybe something like this:<\/p>\n\n\n\n
      \n