Thursday, February 19, 2015

The Memory Half-pipe: Primacy and Recency Effects

We have been doing a lot of memorization in this blog. So why stop now?! Here is another list of words. Do you best to memorize them all. 
  1. Slow
  2. Bridge
  3. Capital
  4. Plug
  5. Ladle
  6. Battery
  7. Keyboard
  8. Insurance
  9. Heater
  10. Shovel
  11. Bread
  12. Goggles

Without looking at the original list, write down all of the words that you remember. This demonstration may fail miserably because, at this point, you are a memory Jedi. The memorization strategies that you have in your arsenal include chunking and placing things in your memory palace

Assume, for a moment, you didn't know about all the various memory hacks we have covered in previous posts. What would you predict in terms of remembering each word? Which words would you be most likely to remember? Which words do you think you would be most likely to forget? If you introspect for a minute, what were some of the things you were doing mentally when trying to memorize the list? 

"There was food in my mouth!" —Mr. Short-term Memory Man

One of the fundamental distinctions in Cognitive Science is the split between long-term and short-term memory. In a previous post, we discussed a very simple model of memory where information enters short-term memory. If it is sufficiently processed, then that information enters long-term memory via a process called encoding. How can we increase the likelihood that information gets encoded?

Probably the most basic method to enhance encoding of information is to rehearse it over and over again. Through repetition you can increase the odds of a specific item getting into long-term memory. Sometimes, if you watch people as they try to memorize a list of words as I challenged you to do earlier, you can see them moving their lips as they rehearse the words. 

Rehearsal is effective at getting items at the top of the list encoded into long-term memory. But as you make your way further into the list, the later items (starting around items 4-9) create a bottleneck because it gets harder for you to loop through the entire list of 4+ items. So the  words in the middle of the list don't get rehearsed as much as the first items and have a harder time making it into long-term memory.

The last part of the list (e.g., items 10-12) also have a hard time making it into long-term memory, but still have a good chance of being recalled. Due to their placement at the end of the list, they can be easily crammed into short-term memory. When it comes time for you to remember the list of words, it's probably easiest for you to quickly list items from the end of the list first, purge those from your short-term memory, and then concentrate on recalling the first part of the list. 


If you plot the probability of recall for each item in the list, you would get the standard serial position curve that you see here in yellow [1]. Items early in the list (slow, bridge, capital) have a high probability of recall. We call that the primacy effect. The items at the end of the list (shovel, bread, goggles) also have a high probability of being recalled. We call that the recency effect because they were the most recent items to cram into memory. Unfortunately, the middle of the list is going to have a low probability of recall.


But wait...there's more!

The serial position curve makes an appearance in most introductory psychology textbooks.  In other words, it is a fairly standard finding. But are there exceptions to the primacy and recency effects? Of course! We've already covered one exception. Memorization strategies can yield perfect recall, which thus negates a primacy or recency effect.

Another exception can occur thanks to an interesting feature of the memory system. Suppose I gave you the list below:
  1. apple
  2. banana
  3. orange
  4. kiwi
  5. strawberry
  6. pineapple
  7. Boeing 747
  8. blueberry
  9. mango
  10. watermelon
  11. lemon
  12. grapefruit

If you were to plot the probability of recall for the above list of items, you would not get the standard serial position curve. Instead, the word that has no business getting a 100% recall rate thanks to its location right in the middle of the list, would get recalled 100% of the time: Boeing 747. This pop-out effect (or the Von Restorff effect) demonstrates that the surrounding items in the list can set up a mental context (fruit) in which we have a very easy time remembering things that break out of that context (Boeing 747).


The STEM Connection

What does this mean for education? Does the primacy and recency effects have an implication for how to structure a lesson? It is always risky to extrapolate findings about lists of words to the complexities of a classroom. But there is a general principle at work, and it's one that journalists understand extremely well: Don't bury the lead. When teaching a lesson, don't put the most important information in the middle of the class period. Instead, either open with it or close with it. Also, you might try to think of a way to leverage the pop-out effect. In fact, maybe Monte Python always had an important contribution to education: And now for something completely different!


Share and Enjoy! 

Dr. Bob


For More Information


[1] Peterson, L. R., & Peterson, M. J. (1959). Short-term retention of individual verbal items. Journal of Experimental Psychology, 58(3), 193–198.