iSixSigma

Learning Curve

Six Sigma – iSixSigma Forums Old Forums General Learning Curve

Viewing 11 posts - 1 through 11 (of 11 total)
  • Author
    Posts
  • #46713

    vidya Kulkarni
    Member

    Dear all,
    I want to put competency levels of employees and thier updation in learning curve.How would the curve look like
     
     
     

    0
    #154715

    Rpaul
    Member

    The curve would generally tend to be somewhat like a bell curve (without the tail areas). This is assuming each curve being for one competency only and not an overall competency index.
    The increase in competency would be slow initially (gradual upward curve), start getting higher and faster, would stabilize and grow proportionately over a period of time, subsequently there would a dip in marginal increments and the curve would start flattening; and eventually the curve would go straight (parallel to X axis) and probably would also dipping downwards (slowly to start, and accelerate later) in the long run. This would apply to most practical skills and not the behavioral indicators (they would not tend to dip down)
    This would be due to the Laws of Marginal Utility where every competency would have an optimum shelf life, after which the efficacy of the competency would drop, and so would the values of the same.
    An overall competency curve shold go up and start stabalizing. It would not drop down unless there is a special cause.
    Again, this is a general prediction and not an absolute one. Exceptions will accrue.

    0
    #154720

    Mikel
    Member

    Wrong

    0
    #154721

    BritW
    Participant

    Wrong.  The initial explanation of the curve is accurate until you reach the top. The flattening will continue to be flat or slightly upward with experience.  As long as the competency doesn’t change, the curve will increase sharply and then flatten – raising slightly.  If the competency does change, then the learning curve will start over from the point of the change.
    The curve will not dip back down unless the person does not use the competency for an expended period of time.

    0
    #154722

    Rpaul
    Member

    “The increase in competency would be slow initially (gradual upward curve), start getting higher and faster, would stabilize and grow proportionately over a period of time, subsequently there would a dip in marginal increments and the curve would start flattening; and eventually the curve would go straight (parallel to X axis) and probably would also dipping downwards (slowly to start, and accelerate later) in the long run.”
    Your point is correct and is a possible factor affecting dip, and exactly why I included “probably”.
    I should have mentioned the reason as well, but thanks for pointing it out anyway!!
     

    0
    #154723

    Vidyadhar
    Member

    The learning curve plotted against competencies will depend on Various factors, level of competency, training for acquiring a particular level, method of testing the learnings etc. Under Ideal conditions the learning curve should show a gradual increase upto a certain level and then move as a straight line, provided you have mechanisms for constant review of the learnings.
     

    0
    #154724

    Vidyadhar
    Member

    Instead of arguing what is the correct curve pattern, the best thing to do would be plott the evaluated performances against time. That would Ideally give u indicators as to where the problem is . Ideally curve should move up gradually & then flatten… if any aberrations are observed, thats an indicator which will tell u something is wrong with the process of evaluating & building competencies

    0
    #154728

    BTDT
    Participant

    Vigya:We have plotted these distributions for a number of applications. We have seen a number of different distributions, but in general:- Output in terms of quantity will start low and grow to a maximum after the ‘learning’ is completed. The maximum level and time to reach maximum level will differ for each operator depending on other factors.- Time to complete a task will start high and decrease to a minimum after the ‘learning’ is completed. This curve is the inverse of the first curve above and will differ for each operator for the same reasons as above.On a related topic we found the time to complete a call in a call centre would follow a log-normal distribution, but the shape of the curve would depend on the experience of the operator. We could prove this using Levene’s test for variation and t-tests on the transformed data for log-mean. The non-parametric test of medians failed to detect a difference.Cheers, BTDT

    0
    #154729

    BritW
    Participant

    Not to belabor the point – but the contest I had with your statement was that there is no probably (especially the part about it dipping and accelerating).  If I graphed this and saw a drop after the curve flattened, then I would be driven to believe there was a significant process change (or blunt trauma to the person because they lost their competence).

    0
    #154738

    CT
    Participant

    BTDT, nice to see your still here keeping the piece.

    0
    #154741

    BTDT
    Participant

    CT:… my piece of what?Cheers, BTDT

    0
Viewing 11 posts - 1 through 11 (of 11 total)

The forum ‘General’ is closed to new topics and replies.