Good stuff, lots of relevant, current references:

A Stability Bias in Human Memory: Overestimating Remembering and Underestimating Learning

Abstract:

The dynamics of human memory are complex and often unintuitive, but certain features--such as the fact that studying results in learning--seem like common knowledge. In 12 experiments, however, participants who were told they would be allowed to study a list of word pairs between 1 and 4 times and then take a cued-recall test predicted little or no learning across trials, notwithstanding their large increases in actual learning. When queried directly, the participants espoused the belief that studying results in learning, but they showed little evidence of that belief in the actual task. These findings, when combined with A. Koriat, R. A. Bjork, L. Sheffer, and S. K. Bar's (2004) research on judgments of forgetting, suggest a stability bias in human memory--that is, a tendency to assume that the accessibility of one's memories will remain relatively stable over time rather than benefiting from future learning or suffering from future forgetting.

- via: "Their results led the researchers to the suggestion that people are underconfident in their learning abilities and overconfident in their memories. That is, people failed to predict that they would be able to remember more words after studying more -- although in reality, they learned far more -- instead basing their predictions on current memory. Kornell and Bjork call this a "stability bias" in memory."

See page 17 for 'Practical Implications' and other comments...

Wow, that sounds horribly flawed. I imagine it something like this:

*Researcher gives highschool student a hammer, then points to the stack of 2x4s and sack of nails.*

*Student proceeds to attempt to nail 1 board to the other. It is crooked and offcenter.*

*Researcher concludes that using a hammer is not an effective way to secure 1 board to another.*

Seriously? 1-4 times? I wouldn't remember anything either. And yet every single one of us has used "studying" to learn the majority of our base knowledge, since every school in the world does it that way.

The point made here is not that studying is ineffective. It's that people underestimate its effectiveness.

People will think, oh I remember that fact, I know it, no need to study it again and then forget it.

The correct thing to do would probably be to put that fact in a SRS before you forget it. But we naturally have a hard time believing we're going to forget things that easily.

Awesome, more scientific support that it's dumb to remove stuff from your SRS just because it "feels easy".

Great, just when started to allow myself to use the "easy" button.

Although, I do understand. This is pretty much the reason I took off the "Next Review in:" thing above the buttons.

Measuring recall is easy. This is what you do in Anki on every review.
How do you measure learning?

What the hell is everyone talking about? The study doesn't say any of that stuff at all (except what nest0r said)... it's about learners' confidence in their knowledge, memory and learning ability, and the importance of how those confidences interact.

Tobberoth Wrote:Awesome, more scientific support that it's dumb to remove stuff from your SRS just because it "feels easy".

Nope...

Blahah Wrote:What the hell is everyone talking about? The study doesn't say any of that stuff at all (except what nest0r said)... it's about learners' confidence in their knowledge, memory and learning ability, and the importance of how those confidences interact.

http://www.phdcomics.com/comics.php?f=1174

iSoron Wrote:

Blahah Wrote:What the hell is everyone talking about? The study doesn't say any of that stuff at all (except what nest0r said)... it's about learners' confidence in their knowledge, memory and learning ability, and the importance of how those confidences interact.

http://www.phdcomics.com/comics.php?f=1174

They got the correlation/cause thing down, hehe.

Blahah Wrote:

Tobberoth Wrote:Awesome, more scientific support that it's dumb to remove stuff from your SRS just because it "feels easy".

Nope...

"The difficulty of such decisions is illustrated by a study (Kornell & Bjork, 2008b) in which participants had to decide whether they wanted to leave a given flashcard in a stack of flashcards so that they could encounter it in the future, or drop it from future study and test trials. Being in control of the decision had a negative impact on participants’ learnin —that is, they did better when they were not allowed to drop any cards—because they underestimated the value of future study and test trials, especially on difficult items, and therefore dropped them too quickly."

Yes.

You can find that 2008b article (costs of dropping flashcards) as well as a paper on spaced flashcards here: http://www.williams.edu/Psychology/Facul.../pubs.html

Their idea of spacing is somewhat different from what we do in the SRS, however. Do factor that into your interpretation of these biases. (Bonus: I like the bit in 2008a about spacing benefiting induction. ;p)

Serendipity: They're part of the growing number of researchers looking into animal metacognition.

Edit: Someone email them and encourage them to use Anki in their research. ;p