Doubt on Sampling
Six Sigma – iSixSigma › Forums › Old Forums › General › Doubt on Sampling
 This topic has 37 replies, 14 voices, and was last updated 18 years, 5 months ago by PB.

AuthorPosts

June 27, 2004 at 5:34 am #35983
BecOMingParticipant@BecOMing Include @BecOMing in your post and this person will
be notified via email.I have an hypothesis.The religion has an influence in credit card purchase of an individual.I live in a city of Inida where there are 3 major religions:Hindusim, Christianity and Muslim. I am trying to do “sampling” with each of these 3 sects and trying to validate my hypothesis with “Chisquare”. My nullhypothesis is ‘religion has no impact’ and my alternate hypothesis is ‘ particular religion will have greater influence in buyer’s mind’.
My doubt is how many samples will I have to take in each of the sects.The total population of Chennai is 4216000.The majority of them is Hindu and the minority is Muslim and Christians. How do I go about estimating the number of samples required?
Regards,
BecOMing.0June 27, 2004 at 3:46 pm #102452
Ken FeldmanParticipant@Darth Include @Darth in your post and this person will
be notified via email.Interesting application. Some things to consider before computing sample size:
What confidence level do you want to have?
What precision level do you want to have?
Do you have any knowledge as to the variation in the population?
These are needed for sample size calculation. Once you have the appropriate sample size you will probably want to sample in proportion to what the different religions represent in the population. These will want to be randomly selected. For chisquare you will need at least 5 responses per cell to do the calculations. If I assume 95% confidence, +/ 5% precision and no knowledge of the population responses you get a total sample size of about 385. Randomly sample from each religion, in proportion to what religion represents.0June 27, 2004 at 3:55 pm #102453
V. LaxmananMember@V.Laxmanan Include @V.Laxmanan in your post and this person will
be notified via email.Dear Becoming:
I have a suggestion for you. Include some Muslim majority towns (there are some in Kerala, and you will also find some in UP), and Christian majority towns (again you should some town in Kerala). Then I have a plan to analyze the hypothesis for you.
I do this differently using what I call a work function not that I have anything against the Chisquare methods, but the new methods seems to yield some interesting insights. Regards.
Laxman
0June 27, 2004 at 4:24 pm #102454V.,As a thought, since many are coming onto this forum asking for assistance with daytoday problems using proven Six Sigma analytics and you are offering an alternative approach work function. Why dont you answer their questions using a parallel approach, first by going ahead and applying routine Six Sigma analytics, e.g., Chisquare, regression analysis, ANOVA/ANCOVA, etc., explaining why you chose the method and what the analytical results are, and then, using a parallel path show in detail the work function application and your results performing a subsequent correlation analysis between the results of both methods describing the plusses and minuses of both methodologies. Whether in pure scientific research or routine engineering problem solving data speaks and rules the day. Or it should anyway if we are as data driven as we profess to be. I, for one, would applaud your stepping to the plate and performing an open headtohead comparison. Pick a good question and provide a good analysis. Thanks in advance for your positive response and analytical work. Frank
0June 27, 2004 at 4:59 pm #102455
V. LaxmananMember@V.Laxmanan Include @V.Laxmanan in your post and this person will
be notified via email.Dear Frank:
How wonderful to hear this from you!
If you check the references that I have cited in my posts, I have just finished writing a book on Six Sigma and Einstein’s Work Function. Here I have discussed how the work function can be used to interpret the most important problem that lies at the very foundation of statistics – the results of a coin tossing experiment. Is the probability of getting a head equal to 0.50 in a real world experiment?
The Nobel laureate Richard Feynman (this is for the nonphysicists, not you, Frank) gives actual data from 3000 coin tosses in his widely read Lectures on Physics (Chapter 6, vol. 1). I have analyzed these results using the work function. I have also analyzed the drawings from the Shewhart normal bowl using the work function. These lead us to a new conception of how to determine probability of the occurrence of an event.
Needless to say, new insights can be gained via the idea of a work function. We can understand, for example, why the probability of obtaining a head keeps changing as more and more tosses are included. Likewise, with the Shewhart normal bowl.
However, I am too “biased” with my work function idea.
What I would like to suggest is that an expert statistician and I work on the same data, and then we compare notes.
In view of the interest that has been generated (I thank everyone who has responded and encouraged, and also those who have been a bit skeptical), I was planning to post my solution to a problem on how to develop a control chart from a classical textbook on Statistical Quality Control. The textbook addresses the approach that everyone is familiar with. You can see what the work function does.
I would also love to get some actual data from more recent observations that are being made evryday by Six Sigma practitioners. This is why I came to this forum and looks like we are going to have a lot of fun soon.
Again, I cannot thank you enough for your suggestion. Enjoy the rest of the Sunday. We should talk about fractals too some day and how it applies to Six Sigma – yes, it does! Regards.
Laxman0June 27, 2004 at 7:40 pm #102456
Chaos NutParticipant@ChaosNut Include @ChaosNut in your post and this person will
be notified via email.V. Laxmanan: Have you read Mikel Harry’s work on the
use of Chaos Theory and Fractal Geometry in Six Sigma
work and the field of engineering? It sounds like the two
of you share some common ground in your thinking. His
book (listed on this web site) does a nice job of
exemplifying how fractals can be employed to emulate
the occurrence of assignable causes during the course of
Monte Carlo simulation. According to Harry this makes
the simulation much more realistic, much like you are
saying. But this type of thinking is so far out outside our
box, most practioners don’t take the time to understand it.
My gut tells me we (as practioners) might be missing
some really good thinking. What is your opinion of
Harry’s approach?0June 27, 2004 at 10:01 pm #102458
V. LaxmananMember@V.Laxmanan Include @V.Laxmanan in your post and this person will
be notified via email.Dear Chaos nut:
I agree with you that chaos theory will help and there may be much to be gained in applying it to Six Sigma and other fields of engineering.
Can chaos theory predict the probability of occurence of heads in a coin tossing experiment? This is where we must begin, in my opinion. And here I already see that there is a work function that we must take into account. The probability of getting a head keep varying as the number of tosses increases and I have been able to explain this (or convince myself) that the work function is the reason for this variation..
If good simulations can take this nagging nonzero c in the equation y = hx + c that creeps up in every real world problem outside physics (where we force c to take the value 0, and with justification), we will make progress.
In all of physics, there is no equation with a nonzero c. Only Einstein’s photoelectric law has the nonzero c. We got rid of the nonzero c in Charles’ law by invoking the absolute zero temperature. But, we cannot get rid of c = – W in the photoelectric law. Even Planck managed to get rid of it when he develop quantum theory, by postulating that there is a ground state, or zero energy state, and so E = n(hf). But he defines temperature by combining the first and the second laws of thermodynamics. To get this solution, he ignores the work done by a body receiving, or rejecting, heat at some temperature T. Hence, T = dE/dS where dE is the change in energy and dS is the change in entropy. This is true only if work done is zero, exactly.
Einstein recognized that work is done when the photon ejects the electron. This cannot be neglected. Mother nature told him that with the cutoff frequency! How simple, once it is conceived!
Hence, K = hf – W. This is the only law of physics where we must confront this nonzero c.
But, outside physics, when we generalize the idea of work, we have forced to confront this nonzero c in every problem. But as with Einstein’s law, only experiments can tell us what W is. If chaos theory can help us simulate how W evolves (i.e. how c evolves in y = hx + c0 we can definitely gain deep insights.
But, I am not the one who can do this. I can only send emails and have mastery over Microsoft Word, Exce, Powerpoint. Otherwise, I am total computer illiterate who has never come close to any simulations. If what I have suggested is appealing, I look forward to the day when others will take these types of simulations seriously.
For example, I just read an article in the latest issue of Fortune (June 28, 2004), where the Ford Chairman and CEO, and family heir, says that we wants to make more money by selling fewer cars. Guess what, it can be done and simulated, if we understand the meaning of the work function. This is what Bill Ford has to change for the company founded by Henry Ford, early in teh 20th century.
There is now more than one car in every American garage but they are increasingly no longer made in America. The work function has changed and it must be changed again to accomplish Bill Ford’s vision.
When Henry Ford founded his company he did three things. He offered double the wages ($ 5 per day) to his anyone willing to come and work on his assembly lines. He lowered the price of the car that he sold. And he increased profits! Can we do it again?
That’s what I am looking forward to. If work function and/or chaos theory helps, along with Six Sigma intiative, all glory to these ideas. Thanks lot for asking. With my best regards.
Laxman
0June 28, 2004 at 12:41 am #102460
BecOMingParticipant@BecOMing Include @BecOMing in your post and this person will
be notified via email.Hi V,
Your post is quite inquisitive.But, as Frank said, can you show us how your approach could be used for my project.Instead of more theory and explanations, I would appreciate if you can just apply values in formulas and show the sequence of steps you perform to arrive to the result for this problem.
Thanks.0June 28, 2004 at 1:49 am #102461Hi Reigle, we’ve been missing you.
0June 28, 2004 at 1:57 am #102462
LR Tobias Jr.Participant@LRTobiasJr. Include @LRTobiasJr. in your post and this person will
be notified via email.Stan since its in reference to Reigle, was your posting name from Stab instead of Stan” perhaps a bit of a Freudian slip? A tad hostile there bud.
0June 28, 2004 at 2:06 am #102463V.If the probability of getting heads changes as you do more coin tosses, then you are in the wrong business. You could make a killing in Las Vegas!Seriously, for all practical purposes the odds better remain about exactly 50/50. Slight variations in the coin itself may make it come up one side more than the other. If you can learn to flip the coin an exact number of times before catching it you could change the odds. If the coin gets worn after millions of flips, the odds could slowly change. But in general, there are so many random factors that there is no way to predict the outcome.The theory you use to predict the next coin toss can’t control the physical act of tossing a coin.Tim F.
0June 28, 2004 at 2:10 am #102464
Chaos NutParticipant@ChaosNut Include @ChaosNut in your post and this person will
be notified via email.One of the main problems with inferential statistics is the
assumption of a static universe. For example, a simple
Xbar and R chart assumes the universe mean and
variance is static over the period of time the chart was
constructed, but we know (in reality) that a universe mean
and variance can be quite dynamic over a relatively short
period of time. If the universe parameters are truly
dynamic, then an Xbar and R chart will ultimately reveal
Type I and Type II errors at a rate greater than the
statistical equations will account for. Another way of
saying all of this is to consider the relation Y=f(X). If X is a
set of causal variables, classic statistics assumes that f
is a constant over the period Y is sampled. If f is a
constant, classic statistics can be applied to sample the
momentary setting of X in an effort to construct a statistical
model in the form of a polynomial. But if the function f is
also a variable, the model loses validity and falls apart.
Perhaps this might be the biggest contribution of Chaos
Theory and Fractal Geometry. By the way, will you please
address my question directly without a lot of discussion
about Einstein, electrons, quantum mechanics and other
such things. Direct answers would be greatly appreciated
and show your commitment toward us applicationminded
practitioners. Once we understand the terms of your
simple work function equation, it will be easier for us to
evaluate its useage and share the same insights you
profess to have.0June 28, 2004 at 2:43 am #102465
Chaos NutParticipant@ChaosNut Include @ChaosNut in your post and this person will
be notified via email.V. Imagine a simple species whose population in one
generation depends only on its population in the previous
generation in two ways. If there are more potential
parents there will be more offspring in the next
generation, but if there are too many in one generation
they each may not get enough nourishment to reproduce.
Also to make things simple, let’s set the units that we use
for talking about the population so that 1 is the absolute
maximum the particular environment can hold. So, in the
ith generation the population xi depends on the
population of the preceding generation xi1 according to
some equation. Probably the simpliest function that fits
the description is an inverted parabola: xi= kxi1(1xi1),
where k is some constant. This equation is very simple,
but it is nonlinear (when multiplied we observe xi=kxi1 –
kxi12). For some values of k, most starting values for the
population, x0, will eventually lead to a single point
(depending on k and not on x0). For other values of k
most starting values for the the population will lead to
oscillating or cyclical values for the population (and the
cycles can be quite long). But for other values of k,
starting values for x0 don’t necessarily converge on any
repeating cycle and the population fluctuates in a way
that is neither cyclical nor random. When this happens, no
difference in starting x is so small that it might not make a
big difference. When a system behaves that way it is
chaotic. Complexity theory (or the study of complex
systems) is really about how a system which is
complicated (usually by having many interactions) can
lead to surprising patterns when the system is looked at
as a whole. For example, each of the billions of water
molecules does its own thing when it joins up with others
as it freezes to others, given some constraints on what
each of them can do, something recognizably snow flake
shaped can emerge. Complexity theory is about how the
interaction of billions of individual entities can lead to
something that appears designed or displaying an overall
systems level pattern.0June 28, 2004 at 3:14 am #102466Since “b” is next to “n” on the keyboard – I’m sure that was the slip.
0June 28, 2004 at 3:19 am #102467
Chaos NutParticipant@ChaosNut Include @ChaosNut in your post and this person will
be notified via email.V. The reason I asked you about Mikel Harrys equations
is the he seems to have successfully demonstrated that a
stable system of random causes can be used to emulate
the influence of nonrandom events, thereby resulting in a
pattern of autocorrelation where the AR1 time series
model is statistically significant. Based on this, he seems
to be saying that a series of random events can be
randomly interacted in such a way that a nonrandom
pattern can be forced during the course of simulation.
Obviously this has huge implications for the field of
engineering design (the designer can emulate all of the
dynamic nuisances of a process, not relying on the
assumption of a static universe). The ability to inject
nonrandom events (in a random manner) using a series
of random numbers might better allow us to simulate
many types of real world conditions we know will occur
over time. It seems he is using a static system of random
causation to induce a dynamic system of nonrandom
events that are overlayed on the original simulation. By
way of his equations, he is able to constrain the limits of
nonrandom perturbations to any predetermined level.
This seems to allow him to set up a process simulation
that factors the influence of certain assignable causes
(like tool wear, shift changes and so on). In the coin toss
example, this is much akin to all of the subtle nonrandom
actions that upset the laws of probability (like coin wear
and so on). I would really appreciate your thinking about
this approach to simulation because it is quite
revolutionary thinking.0June 28, 2004 at 3:24 am #102468V,You asked ” Can chaos theory predict the probability of occurence of heads in a coin tossing experiment?”The answer is yes, but it’s not the answer you want. As anyone who has watched “Jurassic Park” knows, one of the hallmarks of chaos is that small changes in the initial coditions make big changes in the outcome, so that you can’t predict what will happen. For a coin, this means that the small, inevitable variations in the upward speed, rotation and elevation at release, in the air currents, and in the surface where the coin lands makes the outcome completely unpredictable. If it is a fair coin, the odds will be 50/50 for each and every toss, regardless of what theory you use. You also said ” If good simulations can take this nagging nonzero c in the equation y = hx + c that creeps up in every real world problem outside physics (where we force c to take the value 0, and with justification), we will make progress. In all of physics, there is no equation with a nonzero c. Only Einstein’s photoelectric law has the nonzero c.”
Here are several other equations with nonzero constants.Distance traveled at constant velosity:
y = vt + y(0) where y(0) does not have to be zeroLength of rod as a function of temperature:
L = (alpha)T + L(0) where L(0) is the original length of the rodResistance of a metal wire as as function of temperature:
R = (alpha)T + R(0) where the resistance does not go to zero as the temeprature goes to zero.Energy of a quantum mechanical oscillator as a function of quantum number, n:
E = (hf)n + (hf)/2
(i.e. the energy can never be zero, but starts at a minumum of hf/2)If you throw an object away from the earth with a speed greater than the escape velocity, the kinetric energy at the end will be
K = K(0) – W where W is the work done by gravity.I could come up with more if anyone wants. And this last one is analogous to Einstein’s photoelectric effect equation. The “work function” is due to gravity rather than electrostatics but otherwise it is the same concept. So Yes, there are many other exampels, and No, physicists don’t force “c= 0” because in many cases that isn’t how nature always behaves.
Tim F0June 28, 2004 at 3:29 am #102469Chaos Nut – you really are a nut. I am really disgusted with you. You profess to know complexity theory but you plagiarize your postings. I knew that I had read what you just posted somewhere else – and I found it. You lifted it entirely (verbatim) from Energetics as SelfOrganized System: Methodological Aspects, by R. Siugzdaite and S. Novas as, from the Lithuanian Institute of Mathematics and Informatics, published in 2002. You fraud. Before you start claiming to be one of the two authors – be very careful. I know that you are not. I also want V. to show us his analytics, but you, sir, are a fraud. Chaos Nut, indeed. You are no scholar. Get off the forum shyster.
0June 28, 2004 at 3:34 am #102470
radgradParticipant@radgrad Include @radgrad in your post and this person will
be notified via email.Possibly, but its also possible that hes a deranged killer and that he just tipped his hand. Be very afraid Reigle. Take no chances. Stay away from the debate….
0June 28, 2004 at 3:40 am #102471
BecOMingParticipant@BecOMing Include @BecOMing in your post and this person will
be notified via email.Hi V,
I am not very clear about why you would like to include some other cities when it’s not my subject of research.I am only trying to find out with the universe as ‘Chennai’.But, if you need to take any other cities for your equation’ , Pls go ahead. I am curious to know your approach to this problem with your equations.
Thanks.0June 28, 2004 at 3:40 am #102472
Chaos NutParticipant@ChaosNut Include @ChaosNut in your post and this person will
be notified via email.Tim F. Please remember that the world once believed
that human flight was not possible. It was believed the
sound barrier could not be broken and radio waves could
not be transmitted across the ocean. History is full
respected scientists and engineers that regret having said
“it can not be done.” People with a relentless pioneering
spirit and crazy ideas often find unuusal ways to change
the course of human history. They simply “think
differently” about things. I believe there was a time when
the quality profession believed (and published) “six sigma
is ridiculous.” I remember several published articles
proclaiming a variety of reasons why it will never work.
The practice of six sigma is all about reaching for what
most don’t believe can be accomplished. Remember,
there was a time when Einstein was denied a university
position because his theories were considered “absurd.”
Perhaps the poster named “V” is right: six sigma
professionals should first try to collectively “prove” a
rational theory before they tear it apart. If a theory is no
good, the pursuit of its proof will prove fruitless. But if a
theory is right and the first reaction is rejection, the world
might be missing out on something very valuable.0June 28, 2004 at 3:49 am #102473Chaos Nut is Reigle Stewart – he can’t fool anyone no matter how hard he tries.
0June 28, 2004 at 3:59 am #102474
Chaos NutParticipant@ChaosNut Include @ChaosNut in your post and this person will
be notified via email.Frank: You will notice that I never claimed to have written
this post. I merely copied this “excerpt” from the paper
you cited to further support V’s position. I do believe this is
common practice. Since you show a great sensitivity to
this, I will provide full citations in the future, regardless of
context or usage. Such limited citations are not
plagiarism and is commonly done. I simply did not feel
the need to make full citation since I was using it to further
another persons discussion on an informal discussion
board. No where in my post do you read it is my position
that However, your sharp and instantaneous
accusations over such an innocent thing are most
unprofessional. Simply ask and I will tell you (as I am
now doing). Seems you are “laying in weeds” ready to
pounce on anything for any reason.0June 28, 2004 at 4:17 am #102475
Chaos NutParticipant@ChaosNut Include @ChaosNut in your post and this person will
be notified via email.Frank, for your edification: What a distressing contrast
there is between the radiant intelligence of the child and
the feeble mentality of the average adult. Quote from:
Sigmund Freud.0June 28, 2004 at 4:21 am #102476
Chaos NutParticipant@ChaosNut Include @ChaosNut in your post and this person will
be notified via email.Matt: “In seeking wisdom thou art wise; in imagining that
thou hast attained it – thou art a fool.” Quote from Lord
Chesterfield.0June 28, 2004 at 5:05 am #102477
V. LaxmananMember@V.Laxmanan Include @V.Laxmanan in your post and this person will
be notified via email.Dear Tim F:
The theory you use to predict the next coin toss can’t control the physical act of tossing a coin.
That’s what the work function is for!0June 28, 2004 at 5:11 am #102478
V. LaxmananMember@V.Laxmanan Include @V.Laxmanan in your post and this person will
be notified via email.Dear chaos nut:
Direct answers would be greatly appreciated and show your commitment toward us applicationminded practitioners.
………I fully appreciate that and I do want to post one example soon which desribes how the work function is used to deduce the fraction of rejects in a process. I am not a theoretician that all of you seem to think. I had to talk about work function to create a broader appreciation of the nonzero c in y = hx + c. It is one of many ways of explaining what “c” means. May be there are other ways of explaining the significance of “c”.
Believe me, I have rolled up my sleeves and worked in manufacturing plants.
Thanks and please stay tuned.0June 28, 2004 at 5:22 am #102479
V. LaxmananMember@V.Laxmanan Include @V.Laxmanan in your post and this person will
be notified via email.Dear Tim F;
Thanks for providing the other examples.
Sometimes one gets carried away and you should grant a few of those occasionally!
However, except the example of resistance of a metal wire (and the energy of an oscillator), in all the other example one traditionally takes c = 0, mainly because it does not matter.
The point I was making, and which you do support in the final sentence, is that nature does not always behave that way and c is nonzero and we must consider its implications.
But, getting back to the fundamental definition of probability we again use y = hx and take c = 0. That is the reason to talk about the work function, so everyone can begin to see that nature does not obey c = 0 that mathematicians have used for convenience in solving many problems.
One thing is clear to me for all the discussions here. The idea of a work function outside physics is taking root. If I have accomplished that, it matter not what else I do. What I know today is little and there are many here who will be able to do a lot more once we all appreciate the importance of nonzero c – in business, finance, economics, and statistics. Regards.
Laxman0June 28, 2004 at 5:28 am #102481
V. LaxmananMember@V.Laxmanan Include @V.Laxmanan in your post and this person will
be notified via email.Dear chaos nut:
You are asking me questions that I really cannot answer. I have never done simulations, like I mentioned earlier.
I have always developed what we call “analytical” models, not computer models. Computer simulations often involve hidden assumptions. The best course is observations coupled with simple heuristic models. This is pretty much the gist of Einstein’s work that got him the Nobel Prize.
But, I would like, given the opportunity, to work with others who are good doing such simulations and watching from afar what the computer does while I look outside at the same time and do experiments. Regards.0June 28, 2004 at 5:35 am #102482
BecOMingParticipant@BecOMing Include @BecOMing in your post and this person will
be notified via email.Hi V,
Looking forward for your nw approach in solving my problem.
Thanks.0June 28, 2004 at 11:41 am #102503
GabrielParticipant@Gabriel Include @Gabriel in your post and this person will
be notified via email.Ok, that’s enough.
I don’t want to say you are wrong, but I will say that I find it hard to belive. I think that unless you can control every significant factor in the “toss a coin” process, you cannot predict the outcome. And I also think that you cannot control those factors except, maybe, in very tight laboratory conditions. But I accept that I can be biased by preconceptions.
So, lets stop talking about the theory of the photoelectric effect, the escape velocity, and the resonator, and let’s go back to Earth.
I have a few coins in my pocket. Tell me what data you need to show us how your theory works with a real life example, and I will provide it.0June 28, 2004 at 12:25 pm #102510
V. LaxmananMember@V.Laxmanan Include @V.Laxmanan in your post and this person will
be notified via email.Dear Gabriel:
Thank you. I am honored since I know you are one of the gurus here.
I have seen some your explanations on basic statistics questions and was wondering when you will jump in when we are having so much fun.
I have submitted a document for downloading which should be posted hopefully today. Here I have discussed my view of a problem analyzed in textbooks on Statistical Quality Control. I would be happy to discuss this and other such real world issues here on earth.
The experiments I am talking about are being done everyday. All we need to do is take the data and analyze them carefully. By using purely mathematical logic we can make a case for defects y increasing as opportunities x increase. A maximum point is reached and then defects start decreasing. Now, can we obtain “good data” that can be analyzed to test this model, which is based on a generalization of what we call Planck’s radiation law. This the “data” and “experiments” I have talking about.
A table of defects and opportunities just like we find large volume of tables on profits, revenues, assets, share holder equity, price per share, and so on for corporations large and small.
The photoelectric law is simply a very nice way to understand the significance of the general law y = hx + c. It is a special case of the function y = F(x) I have suggested as a generalization of Planck’s law. We can take this literally or metaphorically.
Anyway, I am talking to Gabriel, who is either God or close to God and talks to God. Regards.
Laxman0June 28, 2004 at 12:30 pm #102511
Robert ButlerParticipant@rbutler Include @rbutler in your post and this person will
be notified via email.Your problem falls into the category of questions about stratified random sampling. In your post you indicate that the city of interest has a gross imbalance with respect to the population size of the various religions. For such a situation you would probably want to do stratification with proportional allocation of the samples per population type. This approach will give a selfweighting sample which makes life a lot easier when you have to start running your calculations. There are a number of questions you need to answer before you can develop your sample size estimates. The book Sampling Techniques by Cochran discusses these questions and gives you the mathematics needed to do your sample size calculations. Chapter 5 of the Third Edition is titled “Stratified Random Sampling” and Chapter 5A is titled “Further Aspects of Stratified Sampling”. Those two chapters should provide the information you need to answer your question.
0June 28, 2004 at 12:42 pm #102512Chaos Nut and you now quote Freud? Are you also a Freudian Nut? At least you now provide citations and dont, by their absence, lay claim to wisdom that you dont possess. And, now that I am duly edified, your point in this posting was?
Please, if you post in the future, keep the same forum name because others could then use EVE (Essay Verification Engine) to help you assess the originality of your thoughts as a service to both you and the forum because you appear to have some difficulty there. Frank0June 28, 2004 at 12:56 pm #102514
V. LaxmananMember@V.Laxmanan Include @V.Laxmanan in your post and this person will
be notified via email.Dear Becoming:
I am not really doing anything to solve the problem you posed. I only made a suggestion that perhaps you consider broadening your sample and doing what Robert Butler has mentioned separately in his post.
The more information you can gather the more insights can be gained. Then we rigorously test the hypothesis that you are posing. Regards.
Laxman0June 28, 2004 at 1:27 pm #102518
GabrielParticipant@Gabriel Include @Gabriel in your post and this person will
be notified via email.I saw the thread about you posting the documents after writing the post you have just answered. I will read your documents, of course, and comment on them if I have something to say.
Now, all that stuff about me being a guru, God, etc… It is just a bit of humor, or are you being ironic, or you really meant it?
If it is just a bit of humor, it is Ok.
If it is irony… Maybe the same lack of proficiency in English (just a second language for me) that makes me hard to understand your intyentions made me write the previous port in some way that you found it offensive or something like that. If that was the case, belive me it was not my intention. I honestly meant that I found it hard to belive that you could predict the outcome of a coin better than the statistical way we are used to (i.e. a CI on the occurrence of “heads”), but that “I find hard to belive” is a limitation fromn my side. Not yours. I honestly said that I could be biased. No irony was intneded. And for the theory vs Earth part, it is because I found your theoretical explanations hard to understand (again, may fault) and I thought that a real life case would make it easier.
Now, if you really meant it, you are plain wrong. I am not a guru, not even close. I just try to help when I can (what, as you said, it is on basic statistic questions most of the times), and to express my opinion, understanding or poin of view on cetrain debates. That’s all. I don’t think I am a guru or close to God and I think I never said something in this forum that could lead someone to think I meant that.0June 28, 2004 at 1:42 pm #102524Gabriel – you have more than earned your stripes in contributing to the forum. You have nothing to defend. Just keep moving forward and positively instructing us. V. has yet to earn his forum stripes. Lets give him a chance. We are all anxiously waiting his analytical input after many days of buildup and frequently mentioning Einstein et al. And I feel relatively certain, working in a scientific arena myself, that if I cant understand it once presented, I can get help. Im actually all goosebumpy waiting to see the work function solution.
0June 28, 2004 at 3:38 pm #102538
V. LaxmananMember@V.Laxmanan Include @V.Laxmanan in your post and this person will
be notified via email.Dear Gabriel:
If it is just a bit of humor, it is Ok.
If it is irony… Maybe the same lack of proficiency in English (just a second language for me) that makes me hard to understand your intentions..
Nothing more than just humor, and also a great deal of admiration really for the insights that I see you provide to those who ask fundamental questions in statistics.
I am also one of them – trying to learn what a coin toss experiment tells me in a real world situation, or what is the probability of occurrence of a defect in a process where we want achieve Six Sigma levels of efficiency.
I hope this clarifies what I tried to say in my response. I look forward to discussing statistics with you some day, and learning a lot as well. Regards.
Laxman0June 28, 2004 at 5:33 pm #102557Hi V,
In my earlier post regarding the golf data (Goosen vs. Mickelson) you had developed a y = hx + c function for both golfers. However, you needed some preliminary data to predict (using the Work Function theory) the outomce of the results. You had mentioned that you do need this data (because without which you can not build the model for that golfer). Then, for the coin toss, if I supplied you some data and you build the model (y = hx + c) for this data, will it work for Gabriel’s data or you would need to build a separate function model for his data?
If I build a model and it is working for my process, this model is constantly looking at ‘live’ data and adapting to that new data. I may be able to see that the ‘outcome’ (or ‘output’) keeps changing but what in the ‘input’ has changed, I will not be able to get answer from the model. I may land up using SS for that.
Could you please clarify this further?
Thanks.
Baxi
0 
AuthorPosts
The forum ‘General’ is closed to new topics and replies.