# The Learning Curve

Discussion in 'Human Science' started by FlowerPower, Oct 30, 2001.

1. ### FlowerPowerRegistered Member

Messages:
21
I may be described as anal retentive for bringing this up, but I have noticed for years now that people commonly use the term "steep learning curve" to describe something that takes a long time to learn. Does anybody here agree with me that this is a misnomer?

Here’s my case:

The learning curve is the graph you get when you plot proficiency vs time spent mastering some learned behavior. Now, where I went to school I was taught that when you graph anything that has to do with time, time always is assigned to the x-axis. This means that an easily learned task has a learning curve that is steep, i.e. proficiency increases quickly with time. Conversely, with a task that is difficult to learn proficiency will increase slowly with time- producing a shallow learning curve.

Does anybody care?

3. ### StryderKeeper of "good" ideas.Valued Senior Member

Messages:
13,105

Okay so the above diagram probably isn't the best you've seen of a Learning curve, but I tend to like it.

To put it in a way you can understand, lets say that the top of the Y Axis represents knowing everything about a subject, and that X Axis represents the time it takes to learn.

At first the idea is you start with easy material, that act like the foundation blocks to learning. Take for instance your not going to make much sense of Astrophysics calculations unless you've done the arithmetic that leads up to your understanding of how the calculation works and if the calculation "Looks" correct.

Such things as a learning curve are usually used in Self-assessment to gain an understanding of how many modules a person has done, how many they have left to do, and what level they need to be at to be capable of doing those modules.

I did apply an evolution curve to the finds of ancient artifacts that man used (Kind of every invention), and you notice that if you go from the dateof the first man (well the one that is classed as the olderest) to the present date, this last 2000 years, is like the steep climb arriving at it's pinicle. Of course with evolution we continue, and the internet has increased our curve further, for now we have access to information that before we would of had to have tracked down for ourselves, or waited to be returned to a library system.

I would say that Cybernetic Mnuemonics (Memory implants) would probably increase our evolution curve even further. Imagine they work on a lower level of communication to mobile phones and that you could dial up the information necessary, when your asked a question or need to adapt to a situation.

Of course thats a question of where we go with technology in the future.

5. ### Pollux VRa Bless AmericaRegistered Senior Member

Messages:
6,495
I think the learning curve is just part of our culture and just a word that someone made up that meant something that wasn't its definition, like hot, cool, sexy (maybe), I dunno.

But when you, flowerpower said, "does anybody care?" it reminded me of this joke I made in class. This very attractive girl kept complaining about something but, of course no one cared, and this paper phone just happened to be nearby and I picked it up, dialed zero, and asked the operator: "Excuse me, but could you find someone who cares?" She hates my guts now, and it's a total bummer because she's REALLY hot (no pun intended).

7. ### FlowerPowerRegistered Member

Messages:
21
Hi Stryder- thanks for the reply

I have a question for you concerning your graph:

#1- If "amount learned" is on the y axis, and "time" is on the x-axis and is linear in scale (where every pixel counts for a given unit of time), which side of the graph has the highest rate of learning dy/dx (=slope, or amount learned/time)- the right half or the left half?

#2- Which side of the graph is steeper?

Cheers!

8. ### BowserNamasteValued Senior Member

Messages:
8,828
<i>"Does anybody here agree with me that this is a misnomer?"</i>

I would agree with you.

9. ### StryderKeeper of "good" ideas.Valued Senior Member

Messages:
13,105
Quite simply the curve at the Lefthand side is when you have just started learning and the right hand side peaks at a great height when you have learnt something fully.

If you want to know where the steepest part of the curve actually is, it's in the middle.

10. ### machaonRegistered Senior Member

Messages:
733
Learning curves.

Learning and intelligence are way over rated. So what if universal mysteries unfold in front of you as you are lifted to a higher plane of insight. So what if your mind is an island in a sea of knowledge that laps your shores with waves of levity. You are still going to die. You still need to get laid.

11. ### my_notebookwild hamsterRegistered Senior Member

Messages:
31
************
Quite simply the curve at the Lefthand side is when you have just started learning and the right hand side peaks at a great height when you have learnt something fully.

If you want to know where the steepest part of the curve actually is, it's in the middle.

*************

Quite simply, stryderunknown, you are wrong.

The left side of the chart shows less material learned over a given increment of time than the right side.

The misnomer stems from the fact that people associate the word 'steep' with difficult. But as flower power points out, a steep learning curve is one where the rate of learning increases exponentially over time. It is not a difficult concept.

To respond to one of your original questions, flowerpower, I don't think you are wrong to find fault with the misuse of a term like this. The correct use of language is the foundation of communication, and when language deteriorates, so does our common body of knowledge.

Peace.

12. ### StryderKeeper of "good" ideas.Valued Senior Member

Messages:
13,105
my_notebook

I wasn't wrong I just didn't explain it in full.

I think what was being asked was where is the most knowledge and least time, The steepest part of "Learning something".

I pointed to the middle, as at that point you are learning far more than what you need to learn by the end (as by then you should be an authority on a given subject).

Thats why it seemed a weird answer purely because I didn't unfold it into full english.

13. ### FlowerPowerRegistered Member

Messages:
21
Let me clarify something-

First of all, when I asked the question “which part of the curve is steeper?”, I was specifically referring to the graphic you supplied. So do me a favor and scroll to the top of this post, look at the graphic again, and answer the question. It is obvious that the steepest part of the graph is on the right side, as my_notebook (?) points out it is getting exponentially steeper. Your answer was wrong, but you may have been answering the question in the context of some other learning curve you had jotted down on a piece of paper in front of you.

However, this argument digresses from my point. My fundamental question is this- do you agree or disagree that the steepest part of a learning curve represents the fastest rate of learning? I think your last post answers this question- you agree, right?

Why do I care? I think my_notebook says it best-

What’s the harm with allowing people to use the term “steep learning curve” to describe something that is hard to learn? It’s harmful because it gives people (and especially children) a false sense of intuition about what a curve is. I have tutored youngsters in math, and graphing is one of the most difficult things to teach them. If they grow up with misnomers and bad examples, they will surely have a handicap in education.

History has shown that a society’s knowledge base rises and falls dramatically. The Greeks had a level of mathematic sophistication that rivaled that of 17th century Europe (So did Muslims in Baghdad! Where did their knowledge go? How far would that knowledge have carried them had it not slipped into obscurity? Could educational apathy lead to the same decline in the present world? I think we should strive towards truth no matter how trivial it seems! No matter how steep the curve!

….Oh wait…

14. ### my_notebookwild hamsterRegistered Senior Member

Messages:
31
I pointed to the middle, as at that point you are learning far more than what you need to learn by the end (as by then you should be an authority on a given subject).

OK, but you are not focusing on the argument. Don't overthink this.

Look at the graphic you sent in. It describes a 'steep' learning curve. On that we agree. But look at it. Here are the 2 points that constitute the argument:

where the (y) axis represents amount learned
and the (x) axis represents time elapsed

1.) Toward the right side of the graph the amount learned (y) over a given period of time (x) is greater than it is on the left.

2.) The steeper the curve, the more (y) increases for a given unit of (x)

.. Therefore a steep learning curve describes something that is learned exponentially faster as time continues, and the steeper the curve, the greater the learning.

Like you said, "you're not going to make much sense of Astrophysics calculations unless you've done the arithmetic ".

Don't take it too hard. We're all wrong sometimes. It's whether or not we admit it that makes us men of us.

Messages:
31
16. ### BowserNamasteValued Senior Member

Messages:
8,828
i think people visualize this when they hear <i>"steep learning curve."</i>

17. ### my_notebookwild hamsterRegistered Senior Member

Messages:
31
Hey Bowser.

Interesting. In that case you seem to have a steady rate of learning, which then plateaus and drops off as time continues. Does that imply to you that there is a bulk of knowledge which may be easily learned in a given area, but once that knowledge is attained, additional knowledge is harder to come by - possibly because there are less secondary resources, such as textbooks or teachers?

It seems to me that it is hard to graph these things out simply because learning is so subjective. How could you measure the learning curve for skiing, for example? For one person it is hard and for someone else they are busitng out helecopters at the end of the first day.

I think that 'learning curve' is, as Shrike points out, just one of those terms that slipped into the common vocabulary unnoticed, and I'll bet dollars to donuts that it was the fault of some corporate trainer who thought it made them sound intelligent and better educated.

It's only a word, I know, but like FlowerPower said it leads to a gross misunderstanding. It *seems* like the right term, just as the sun *seems* to move across the sky every day, but we know better.

We NEED to know better!!

18. ### BowserNamasteValued Senior Member

Messages:
8,828
Let me try to support that graph. I believe it represents the difficulty in obtaining full comprehension of a subject over a period of time--the left side being the start of the study and most difficult stretch of time, the right being the downward slide towards completion of that study.

To be honest, however, if we were to flip Stryder's graph horizontally, we would then have a better representation of the learning process over a given time--work to be accomplished, I suppose.

Last edited: Nov 8, 2001
19. ### my_notebookwild hamsterRegistered Senior Member

Messages:
31
the left side being the start of the study and most difficult stretch of time, the right being the downward slide towards completion of that study

Right, but the thing is that although a steep 'uphill' stretch on the graph *looks* like the "hard" part, and the "downhill" stretch *looks* like the easy part; the opposite is true, because a more vertical slope shows that the person learned more material in the same time. So in fact a steeper section of graph is "easier" and not harder.

People look at the graph and think about climbing a hill. That is why they think a steep learning curve is a hard one, when in fact the opposite is true.

What you are describing is a graph where the x axis represents the time spent learning and the y-axis represents the difficulty of learning. In such a case a "steep" learning curve would be a "hard" one. However, you probably wouldn't want to draw the graph this way because difficulty itself is completely subjective. When you graph amount learned/time you can use objective means to determine how much a person has learned, such as a knowledge test, and even though the curve will differ for each person who is tested, you nonetheless get an accurate graph. And, you can then look at that data to determine a "difficulty rating" or some such thing.

I can't think of a way to objectively measure difficulty during the course of the data-collecting itself, so I am ruling out the possibility of a difficulty/time model. This is, in my mind, the first misunderstanding of the *steep learning curve*. The second is that a learning curve differs only by subject and not by the person who is learning. "Linux has a steeper learning curve than Windows" would be an example of that second misunderstanding.

Help me out flowerpower... am I explaining this correctly?

20. ### StryderKeeper of "good" ideas.Valued Senior Member

Messages:
13,105
The image I showed was one that I had been shown while discussing something else but it involved a learning curve.

What I was discussing was the use of equipment for transversing design work into compressed time packets (Namely 40 years of research could be outputed by a machine in 5 minutes) the problem was simple:

Take the original learning curve diagram that all of you are convinced is wrong, and say that your 100% complete is at the end of the curve on the (Y axis).

Now you take your machine for the time compression, and you assign a "Quantum Leap factor" and you draw a diagonal line between the time period that the machine is to gain knowledge at (somewhere on the curve) to the point on the curve you want to learn.

The understanding was that your machine could learn the point near the (Y axis) at the point near the (X axis), but this then gave the problem that your learning curve ending has now shifted, since the data you were amounting has been moved.

The person I was speaking to asked what you would do then? at the time I just saw a shift in curve, and didn't tally a consideration about a fractal episode and a possible reuse of the equipment after its 100% point.

That is why I used the curve I provided, although the one Bowser has provided is efficent also (It's just like taking Bowsers and folding it in the middle, so one half is at a right angle to the other.)

21. ### BowserNamasteValued Senior Member

Messages:
8,828
Now I'm confused...

22. ### StryderKeeper of "good" ideas.Valued Senior Member

Messages:
13,105
Okay it looks odd, you might of been able to stretch a preportion or two, but the bulge in the middle, is pretty much what I previously pointed out, the place where the most is learnt and not relying as heavily on what is already learnt.

Not a bad job though Bowser for the folding of it

23. ### my_notebookwild hamsterRegistered Senior Member

Messages:
31
Now you take your machine for the time compression, and you assign a "Quantum Leap factor" and you draw a diagonal line between the time period that the machine is to gain knowledge at (somewhere on the curve) to the point on the curve you want to learn.

??

The dog ate my machine for time compression.

I left my Quantum Leap factor at home.

What the hell were we talking about?