In this lecture we'll be talking about long term memory, which is where we store any information that we access more than a few seconds after we're exposed to it. And as you'll recall, not every piece of information that we're exposed to makes it to our long term memory. So there's a lot of information that's presented to our senses that we never perceive in the first place. And a subset of what we perceive makes it into short term memory, where we can operate on it for a short period of time, maybe a few seconds. And a subset of that information that enters our short term memory is actually committed to long term memory. So long term memory contains anything that we remember for more than a few seconds. And for information to become part of our long term memory it has to be copied from our short term memory to our long term memory, and that process of transfer from short term to long term memory is what we call learning. So how does the transfer to long term memory happen? Well, it happens, really, in two primary ways. One is through association where some piece of information that enters our consciousness is associated with things that we already know that makes it easy for us to kind of fit in with the things that we know. And the other way is through repetition so when were presented with something we don't really have any associations for, really the only way that will be committed to our long term memory is by effortfully repeating it to ourselves until we can commit it to memory. So let's take an example of how associative memory works. So our memory is fundamentally associative, all of the concepts that we have in our memory, are associated with other concepts and kind of a network structure that links them together. And if you have a set of associated concept that kind of all fit together like the concepts that are shown here, we refer to that as a schema which is really just a way of talking about a particular network of related concepts. So these are concepts that are all related to the idea of a farm. So you have animals that are on the farm, horse, donkey, cow, and things like tractors, fields, pastures, corn, and so on and so forth. And if we're presented with a new piece of information, let's say, for example, you're presented with information which I realize may not be new for most of you, but that if you have an animal who has one parent who's a horse and one parent who's a donkey, that animal is called a mule. Well, that's a piece of information, if it were new, that would be fairly easy for somebody who had those other concepts to remember because it fits right in. And so you have simple association between the concept of mule and the concept of horse and the concept of donkey. And to that information that's going to be easier and more likely to be committed to long term memory and more likely to be remembered later. In contrast if I presented you with an animal called plzykofwkblu and it looked like this you might not be very inclined to remember this because it wouldn't fit in with things that you currently know and it wouldn't have any place to associate. Now if we don't have already made association for a new piece of information, we're not going to be able to fitting in very easily and we're going to have to put a more effort to remember it. So we engage them in a process of memorization or elaborative rehearsal. Which is basically where we repeat information to ourselves until we've committed it to memory. So this is what happens when we're presented with information like this, call us at 1-895-555-9368. That's going to be very difficult to remember that sequence of numbers because it doesn't have patterns that are easily remembered and it doesn't fit in with information that we already have. Or if you're assigned a user id that's 76277,303, which by the way is the way that a lot of user id's were assigned to people in the early days of internet service providers, before they let people choose their own id's. And this is obviously going to be much harder for people to remember. Much more likely that they'll have to write it down and keep that piece of paper handy whenever they're logging in, then if they get to pick their own name which is meaningful and probably has associations with things that they already know. Memorization is also how we have to learn arbitrary or complex commands such as the things that we need to learn when we interact with unix or a system that requires textual commands. And it's also the technique that we have to use when presented with directions, for example, like go straight and take the second left, then turn right and left again. And after the bridge, turn left, and it will be on your right. That's going to be very difficult to remember, because it doesn't fit in and we don't have anything that we can associate it with. The only way we're going to remember it is either to write it down, or to commit it to memory and repeat it to ourselves. So given these two options for how to memorize or how to commit information to long term memory, which do you think is going to be easier for a user, and which is going to be more likely to be remembered? If you said associative you're right. It's much easier for users to learn and remember things that fit into an associative scheme that they already have. Than it is for them to put in the effort to memorize something that doesn't provide ready-made associations. And the likelihood of remembering something is based on three factors with regards to that information. So one is the strength of association, how closely associated is it with information that the user already has and has already remembered? But also, the recency, how recently were they exposed to the information, and how frequently have they had to recall that information. Another important thing to keep in mind about remembering information is that memories are often primed by stimuli that we receive from the outside world. So we might be triggered by a word or an image or a sound or a smell to recall information that's associated with that, if we've seen that image or heard that word or heard that sound or smelled that before. So we can present information that will fit into users schema by using metaphors, leveraging standards and consistency. And also when we do present information that fits into users' schemas we avoid asking them to memorize stuff and make it more likely that they'll be able to remember it when they need that information later. And another principle that we can apply is that we should prefer recognition over recall whenever possible. So one technique that we can use for presenting information in a way that fits into a user schema is to use metaphors. So one metaphor that many of us are familiar with is the idea of a shopping cart on an e-commerce site like Amazon. And this leverages the metaphor of a physical shopping cart that many people would have experienced in a grocery store for example. Where you can add things to the cart and even take them out before you reach the check out counter. And you only have to pay for things when you actually get to the point of checkout. And so by leveraging the metaphor of a cart, Amazon and other retailers were able to take advantage of the metaphor to explain to users how to use the site in a way that would make sense to them. Another technique that we can use to help users make sense of complicated system, is to take advantage of consistency across different similar products and different parts of the same product. And to take advantage of standards which are similarities across competing products that might be trying to accomplish the same thing. To look at examples of this, you can see that this image up here which is from Google Docs uses a certain menu structure file, edit, insert, format and so forth which is replicated in another one of Google's products, which is Google spreadsheets, right here, or Google Sheets. And it uses the same menu structure. And what this means is that a user who has learned to use one of these systems in picking up the other one will immediately know where to find certain types of operations, like cut, copy, and paste, or save or rename, or those types of things. More over even completely different product, Microsoft Word Online is going to use a very similar menu structure, similar placement of options like formatting, changing fonts, and things like that. To take advantage of the fact that users who've learned one system are going to be able to rapidly transfer that information to another system and not have to learn everything from scratch. So using consistency and standards helps users who have already used a similar system, use the system that you're designing without having to retrain. Wherever possible you want to avoid asking your users to memorize stuff. One place that this comes up is in asking people to choose passwords. Passwords are a fact of life, many systems require them for very good reasons, they need to keep the data secure, they need to make sure that you are who you say you are and over time more site require more complex passwords in order to maintain security. And so here's an example of the kind of security policy that you might see for passwords. Passwords has to have eight characters in length and contain at least three of the four following items, uppercase, lowercase, numbers, and symbols. By adding these constraints you're forcing people to choose potentially more complex passwords that might have fewer associations, and making it more difficult for them to memorize the different passwords that they have. Especially as different sites have different requirements. This could actually create a less secure situation because people have to write these passwords down. They might write them on post-it notes that are kept next to their computer. They might write them in documents that are stored on their hard drives, and things like that. And it creates a difficult situation. The point that I'm trying to make here, is that by asking users to memorize stuff, it makes things more challenging for them, and can cause them to use workarounds, which actually might subvert some of the things that you're trying to accomplish. Another principle to keep in mind is that you want to prioritize recognition over recall, that is people are going to have a much easier time recognizing an option if they've seen it before than recalling it from memory with no prompting. Most graphical user interfaces do this very well by presenting the menu options that are available, the objects that can be operated on in the form of icons, and things like that. But there's still a few places where we're expected to recall things from our memory. One of them is passwords, which we just talked about. Another is when we're searching for things. We have to call those search terms up from our memory to find the things that we want. So a few years back, Google introduced the auto complete feature for searching to facilitate this. So we have to start with our recalled search term. As we type it we can see other suggestions that come from other people's searches. Which might help us fill that out so we can actually recognize the search term that we're interested in from the options once we've started from a base of what it is that we think we're looking for. So using these principles will help us take advantage of our knowledge of long term memory when designing systems that will work better for people. So learning will work better if the learner can fit new information into a schema that they already have. And we can help them do that by using metaphors. By leveraging standards and consistency and through doing this we can avoid asking users to memorize stuff, which will be difficult for them. By preferring recognition over recall, by presenting options to users that they will recognize rather than asking them to remember things from their own memory, we can also help people accomplish tasks and design systems that will work better for them.
