How to remember metric units - Texts - Memory

The magic number 7 and the memory span

Our mental working memory and the memory span
A classic: the magic number 7
The magic 7 on the internet: practical consequences
In reality: 7 chunks?
Conclusions Basic Research and Reality

There is a lot of talk about information overload on the internet and, as we all know, there are good reasons for this. There are dozens of pages on almost every topic with hundreds of links and menus teeming with somewhat vague one-word choices. There are portal pages that are crammed with links and texts in microtext - in short: we have capacity problems. Or rather: our short-term memory, i.e. the working memory of our consciousness, has capacity problems. 
Our mental working memory and the memory span

Now this is precisely one of the best-studied phenomena in memory psychology - the limited capacity of our short-term memory to absorb. As early as the 1950s there were experimental studies that tried to determine the extent of the so-called memory span. This is the amount of information (e.g. numbers) that we can memorize for a few seconds without specifically memorizing, e.g. by speaking it quietly to ourselves and repeating it. Remembering a phone number for a short time is an example of this. A telephone number is: 8 9 0 4 6 5, and you, dear reader, could now easily remember this, go to the phone and dial (but please do not try it!). The 6 numbers "fit" into your memory span or your short-term memory. Your mental working memory - to repeat the analogy with the computer - takes up the numbers and nothing is lost.

The size of this main memory is of course a decisive factor if you want to assess whether the amount of information to be processed or to be memorized can be managed - this is no different with humans than with computers. A PC of the historical series 80286 would feel as competent when attempting to start Windows 2000 as some Internet users when looking at a hit list with 354 hits that is supplied to them by a helpful full-text search engine.

But how can you find out how much information a person can store in their short-term memory on average? Let's do a thought experiment: We read series of numbers to a group of test persons, first two numbers, then three, then four, etc. After each series we interrupt and our test objects have the task of writing them down in reverse order immediately from memory. (An exciting experience.) The notes are collected and we count how many digits are missing or wrong. The amount at which the test subjects begin to forget digits or make mistakes marks the limit of the short-term memory capacity. This will not be 100% exact, but by counting and taking averages we could come to a reasonably exact result.

A classic: the magic number 7

It was a simple principle, and many similar experiments were made in the 1940s-50s. Then in 1956 a Mr. Miller published his famous article entitled ...

"The Magical Number 7, Plus or Minus Two: Some Limits on Our Capacity for Processing Information"Psychological Review, 1956, Volume 63, pages 81-97.
... in which he summarized and compared the results of all these studies. His conclusion: The memory span is 7 plus / minus 2 "chunks".

What is a "chunk"? The best way to demonstrate this is with an example. Let us assume that we read out individual rows of letters of varying lengths to our test subjects again, e.g.

Then we will find out again: at some point you will become insecure and no longer be able to reproduce all letters correctly. That will be about

be the case, here there are 8 letters that have to be kept.

We conclude: humans can store a sequence of 7 letters in their short-term memory without errors. Miller was right about the 7, but where are the chunks? Now a little trick. I just rearrange the 8 letters that have just overwhelmed us:

and it's no problem at all to remember them. What happened? We have formed a single item of "higher order" information (the word) from disjointed pieces of information (letters). This process is called "chunking", and the term "chunk" refers to a unit of information that combines several elements into a single meaning. Miller drew attention to it - and he proved this with the help of the scientific literature - that one cannot memorize seven numbers or letters or words, but seven "chunks". So when one speaks of capacity, one has to abstract from the content. To speak of a memory span for numbers or letters - as our experiment has just shown - does not necessarily make sense.


To make this clear again you can remember 7 relatively simple words as well as 7 letters, so

Not particularly original, but you can see the principle quite well. The 7 chunks (here words) contain 35 letters that we could never remember as disjointed pieces of information. And - you may have guessed it - we can form a "chunk" again, namely a sentence:

Somewhere, of course, you come across a limit when chunking, you can no longer remember seven sentences, seven stories and seven books (that would be nice). Nevertheless: Chunking is one of the most effective strategies that our brain has developed in order to be able to memorize large amounts of information in compressed form. Without this technology we would be lost with a tiny memory that cannot hold more than seven letters or numbers. Chunking is an intelligent service, so anything but a mechanical process. The compression of pure data into something "with meaning" requires knowledge of meanings and the ability to link information.

The magic 7 on the internet: practical consequences

All of this is not particularly exciting, but at least nice and clear (which is unfortunately all too seldom the case in psychology). And Miller's magic number 7 is one of the best-known and safest findings that memory psychology has to offer. It is also of great practical importance, because of course a visitor is not served particularly well on a website with 3 menus, each containing 22 options: They do not fit into his short-term memory at the same time to be compared or evaluated there. Problems arise especially when information is so extensive that it can no longer be displayed at a glance. Then we have to switch our attention back and forth several times, store information temporarily, compare it with others, etc. What can be seen at the same time, on the other hand, does not have to be fully adapted to our capacity limits - the screen can be an external short-term memory. For an ergonomic page design, the spatial distribution of information is decisive: what has to be compared or processed in another way at the same time should be grouped so that it can be seen at a glance.

However, anyone who clicks from one hypertext page to the next will - as the law suggests - only "take" around 7 pieces of information. When you click through dozens of pages, almost everything is completely forgotten that has not been read or observed particularly intensively. Of course, not everything is really forgotten. The memory span is defined as the amount of information that we look at without any special effort and for a short period of time can remember. It is no problem to learn 12, 13 or even 240 words if you have the time and make an effort - although very few web surfers will be ready for it. However, this is no longer a pure short-term memory function, because information that we learn is stored in the long-term memory.

If chunking (the contraction of individual information elements into units of meaning) really results in a multiplication of the processing capacity, isn't it a sensible strategy to make higher-level units of meaning also visible? Indeed it is, as the following example shows:

contact formcontact form
e-mailFeedback questionnaire
Company historyNew products
New productsalphabetical product directory
Feedback questionnaireOrder
Customers / referencesCustomers / references
alphabetical product directoryServices
OrderCompany history

In the unordered list, the information is presented in such a way that the superordinate categories (chunks) cannot be easily recognized. The short-term memory is overwhelmed. In the ordered list, on the other hand, we only have to learn 3 chunks, namely "contact paths" (yellow), "products" (pink), "company information" (light blue), and can subordinate the 9 links to this system.

In reality: 7 chunks?

Miller's magic number is not only prominent in psychology. It is one of the few results of basic research that has found its way into practice. Anyone who has taken a screen design course knows: "How much information can a person keep in short-term memory?" - "Seven." Sure, of course! But, unfortunately, unfortunately: it is wrong, or should we say better: incomplete and a bold assertion. This does not affect the process of chunking or the fundamental validity of the methods that were (and are) used to determine the memory span. If you define and examine it in this way, it actually comprises 7 units. The problem is completely different:

The reality.

Let's do our thought experiment again. This time, however, observe a special person - let's call her Ms. G. - to check the extent of her memory span. But stop! Not in the laboratory, but under combat conditions, in harsh reality. And here Ms. G. has better things to do than memorize numbers or letters or chunks. It has goals, and keeping a goal in mind already costs one of the precious memory spaces.

In very few cases, however, goals are simple and one-dimensional; they can often only be achieved if they are nested in one another Intermediate goals defined and processed.

The subjects in psychological laboratories (where the experiments are done) are usually motivated. You want to intentionally memorize the information and do so as well as possible. Ms. G. does not want one or the other, and that is her right. So:

We're not finished yet: bad ones Whim has been shown to have a negative effect on short-term memory (good mood too, by the way). Ms. G. was bullied again yesterday by this creep - her boss - and that is still on her mind.

And the neighbour's teenage son has turned the stereo up to full again. Techno! That is the purest noiseharassment!

But now it is really getting late, Ms. G. is tired:

And then she always takes her nightcap (cherry liqueur, 40% alcohol)

So now we have reached the point where Ms. G. has to stop dealing with our beautiful Internet pages, because her 7 chunks of comprehensive information processing capacity is down to zero.

But there is more, because - had I not mentioned that yet? - Ms. G. is about to retire.

We are now reaching the zone in which Ms. G. apparently knows less than before after presenting the information that she is supposed to remember. And - I'm saying it straightforwardly and bluntly - it was actually never very bright.

So now we've also deleted the goal and the intermediate goal from above, and that should really be enough.


So we see Miller's magic number 7 melt together considering some of the factors that distinguish everyday life from a laboratory setting. And this applies in general: Whenever performance is measured in standardized conditions with highly motivated test persons, you get results that are "too good" - as if you were walking at the pace of everyday life (in the rain, loaded with plastic bags, etc.) ), wanted to measure in the stadium. Therefore the rule of 7 is not fundamentally wrong, only: From the perspective of the Internet users, it does not indicate an optimum, but an maximum at. If one assumes a value of 3-5 units for the extent of the memory span, one is closer to the real conditions. That is ridiculously little if you have to pack and link 200-300 Internet pages, for example, but we saw above that you can do something about this, for example by intelligently arranging relevant information.

But: there is at all an optimum? Yes. The optimum that can be achieved when presenting information for short-term memory is 1. In other words: It is optimum when exactly the one piece of information is displayed that comes closest to the visitor's goals or - that can be just as important - information that makes it clear that a current goal cannot be achieved. This topic (actions, achieved and not achieved goals while surfing) will be dealt with in the next article in the text section.