Showing posts with label data. Show all posts
Showing posts with label data. Show all posts

Friday, February 07, 2020

Experimenting to Improve Sleep Quality

comments on: Can a Humidifier Help You Sleep Better and Snore Less?

After doing some research I learned that humidifiers have helped folks snore less. So, after some more research, I picked up a slick little ultrasonic humidifier and gave it a try. Now, it’s been less than a week which I know isn’t enough to get too excited about statistically speaking. But one thing is becoming crystal clear…it’s most definitely helping me sleep better.

Interesting post, which includes control charts showing the impressive progress.

"I’m also still trying to figure out what caused the three special cause signals in January." One nice aspect of improvement is sometimes you can make a system improvement that even without knowing the causes of previous problems, the new improvement stops those from happening again. Maybe that won't be the case this time but maybe it will. Health related issues are so touchy that I could imagine it is something like a couple bad factors stacked on top just push things over the limit. So being a bit tired and say too low humidity and you didn't drink quite enough liquid and sleep quality is bad but just 1 or 2 of those and it might be a bit worse but not horrible.

Special cause signals will be more frequent if several factors together amplify each other (and they rarely happen together so those amplified results are rare). What happens is those rare amplified events will be special, outside of the system that generates that regular variation when they act alone but when all that variation lines up just right the result will be outside what is normal (due to the very large change in the result for that special case where the individual factors acting together (amplifying) create a very large change in the result.

Related: Gadgets to Mask Noise and Help You Sleep or Concentrate - Apply Management Improvement Principles to Your Situation - Zeo Personal Sleep Manager - Using Control Chart to Understand Free Throw Shooting Results

Tuesday, October 08, 2019

The Importance of Providing Context for Data

Data in Everyday Life: How Serious is the Student Loan Debt Crisis?

Did you know that 1 in 4 Americans have student loan debt? In the United States, there are more than 44 million people who collectively owe $1.56 trillion in student loans.

I thought 25% seemed too high. Thankfully this post also provides the number, 44 million. The USA population is currently estimated at 330 million. 44 million is closer to 14% than 25%. My guess is that the 25% is of some subset of the USA population (people between certain ages maybe)?

I think a point about the importance of providing the proper context when showing data has also been made in this post. There is certainly a sensible argument for why 25% of certain ages is a more useful figure for understanding how widespread student debt is (rather than saying 14% those in the USA have student debt). But in the case where that decision is made the details should be spelled out.

Related: Operational Definitions and Data Collection - Data Can’t Lie (data can be wrong) - Poorly Stratified Data Leads to Mistakes in Analysis - Understanding Data

Monday, November 19, 2018

Using Control Chart to Understand Free Throw Shooting Results

comment on: Is Andre Drummond Sustaining His Free Throw Improvements?

Nice use of data to understand the system results (and if there is a measurable improvement or not). I still want to see these chronically poor free throw shooters use the pretty clearly better underhand free throw style. I wrote about this previously in Why Do People Fail to Adopt Better Management Methods?

It really is remarkable how much effort is put into so many aspects of gaining small advantages yet a fairly obvious big advantage (underhand free throws continues to be ignored).

Related: Lessons for Managers from the Wisconsin and Duke Basketball Programs - Change Management - Post Change Evaluation and Action - Taking Risks Based on Evidence

Friday, September 14, 2018

Understanding and Using Data: Waffle House Example

Comments on, Why FEMA is Monitoring Waffle House this Weekend

look outside your organization for risk indicators that might help you make better (and faster) decisions, particularly when those risks are activated. Second, that you should explore crowdsourced risk data as a source of up-to-date information.

I agree. Also, just look at data close to where the action is (internal or external).

And don't just look at aggregated data but dig into what individual data points tell you. Aggregated data is very useful but it also can mask meaningful insights available when data is looked at more closely. It isn't a perfect match to Waffle House data but I think the principle is visible in the Waffle House example.

The Waffle House closure data is based on actually closing based existing conditions while warnings and evacuation recommendations are based on predictions about the weather and the impacts those will have on locations. The warnings are necessarily predictions (to be useful for the whole community they need lead times to take action) where the Waffle House has more flexibility and the organization has managed their system to be more capable of adapting to harsh conditions. There is a real similarity with designing a agile software development process that is able to be more flexible and react quicker than old "waterfall" style organizations that have to predict far in advance and adapt slowly as conditions change.

Waffle House closures are a useful data point for conditions on the ground that FEMA can use with other data to make decisions on how to react.

Related: Bad Weather is Part of the Transportation System - Leadership: Taking Action When Others Are Unsure - Lessons on Competition from Mother Nature - All Data is Wrong, Some is Useful

Thursday, May 25, 2017

Interactions Among the Four Fields in Deming's System of Profound Knowledge

Question on Reddit:
I would like to see expansion of Deming's SoPK with more examples of interactions in the four fields of knowledge. Are you aware of any?

My response:

I think this is actually very common, but often it isn't explicitly mentioned. To explain this well would take a fair amount of time. Let me just give 2 quick examples

Distorting the System to meet a target

This certainly is about the interaction of understanding variation (in this case people not understanding data well enough and being mislead), psychology (how people respond to pressure to meet goals), theory of knowledge (not understanding the difference between the proxy value of data and the underlying truth) and systems thinking (how a system is likely to react to meet goals - distorting data and distorting the system, and using simple measures where those things work to get numbers).

Create a System That Lets People Take Pride in Their Work

Appreciating that results will be better when people are doing work they are proud of involves at least appreciation
for a system and psychology.

I think in reality nearly every example involves interactions. We can analytically separate out the one we want to discuss or the one that seems most influential in what we are looking at but in reality it isn't just one.
For example, data issues related to over-reacting to common cause variation is in the "understanding variation" realm. But it is also deeply ingrained in our psychology that we look for special causes. If our psychology was different it is very possible the mistake of "seeing" (and believing) special causes everywhere would not be a problem. But because those 2 area interact in the way they do it is an area of improvement for how we think and manage. By focusing on an understanding of variation we can limit the damage caused by are faulty psychology (seeing special causes where they don't exist - where it is just common causes). And that really integrates theory of knowledge and systems thinking (we chronically over-simplfy and ignore the large system).

Related: 94% Belongs to the System - Encourage Improvement Action by Everyone - Circle of Influence

Thursday, February 09, 2017

Should I be in the Check Phase of PDCA Daily?

Below is my response on closed forum about whether doing the "check" phase of PDCA daily was too often. I expanded on my comments there a bit in this post.

The check/study phase should be reviewing the results of the experiment done in the Do the experiment phase. "Checking" how things are going during the experiment makes sense but that isn't the check/study phase of PDSA .

For example, you don't want to pay no attention during the experiment and then look at the data and discover the data shows obvious signs the operational definitions were not clear, or the process is providing very bad results. So you need to have those doing the experiment paying attention daily.

Remember one key to using the PDSA cycle is to turn through the whole cycle quickly. Daily would be exceptionally quick. Moving through the whole cycle in 2-6 weeks is more normal. Organizations successful using PDSA will quickly turn the cycle 4+ times for a specific effort (often the 2nd, 3rd... times through are much faster than the first time through).

More on how to use the PDSA well:

Saturday, September 03, 2016

How to Improve at Understanding Variation and Using Data to Improve

My comments based on a question on, How to Use Data and Avoid Being Mislead by Data:

Thanks for this post John. This is the part of Deming’s teaching that I often struggle with (understanding variation). I read Wheeler’s book Understanding Variation and it helped me with the concept, but I am challenged trying to apply it where I work. I often am not sure what to measure and if I do, I’m not sure how to measure it. Folks appreciate my burn down charts showing trends, but this is about the best I’ve been able to do. Do you have any recommendations on where I can look to help me get better at this?

Getting better at using data is a bit tricky, so struggling is fairly common.
Probably the easiest thing to do is to stop reacting to normal variation (caused by the system) as if it were special. This isn’t super easy but it is the easiest step. And it does make a big difference even if it doesn’t seem very exciting.

The idea of actually using data properly provides big benefit but it much trickier. Don Wheeler’s book is a great start. Making predictions and evaluating how those predictions turn out is also valuable. And in doing so often (though not always) it will also spur you to collect data. This process of predicting, figuring out what data to use to help do so (and to evaluate the results) and considering the result of the prediction and how well the predictions overall are working can help.

You learn what data is often useful, you experiment with real data and real processes and you learn what needs to improve. If you are at least somewhat close to using data well then just doing it and learning from your experience is very useful. If you are really far off the experience might not help any 🙁
The links in the post above I think provide some useful tips (and the links within the posts they link to…).

More: Measurement and Data Collection


If you don’t have an answer for how you will use the data, once you get it, then you probably shouldn’t waste resources collecting it (and I find there is frequently no plan for using the results).

It isn’t uncommon that the measures you would like to have are just not realistically available or are hard to determine. How to get started in this is one of the tricker pieces in my experience. It is a place where consultants may be very helpful. If that isn’t an option another possibility is just to ask others at your workplace for ideas for metrics (there are issues with this and a big one is that many metrics will more likely to lead you astray than actually help).

This can also be an area where seeing what others are using can be helpful. Because it is hard to think up what are great metric seeing what others are doing may provide insight. Of course, the ideas must be evaluate for whether they would work for you (even if they are right for others they may not be right for you – and many are not really right for others it is just a thing they measure and while they have associated it with good things maybe they are wrong (correlation but not causation]).

Monday, March 14, 2016

William G. Hunter Award (nomination deadline June 30th)

William G. Hunter Award
Nomination Deadline: June 30

Criteria for Selection - The William G. Hunter Award is presented annually in order to encourage the creative development and application of statistical techniques to problem-solving in the quality field. Named in honor of the Statistics Division’s founding chairman, the award recognizes that person (or persons) whose actions most closely mirror Bill Hunter’s strengths, which were as:

  • A Communicator
  • A Consultant
  • An Educator (especially for practitioners)
  • An Innovator
  • An Integrator (of statistics with other disciplines) and
  • An Implementor (who obtained results)

Download Award Criteria and Nomination Form (DOC)

Past awardees include: Gerald Hahn, Brian Joiner, Soren Bisgaard, Christine Anderson-Cook and Bill Hill.

Monday, January 11, 2016

Don't Use Targets as a Management Tool

comments on: Deconstructing Deming XI B – Eliminate numerical goals for management


I agree with the comments that targets are unwise. There is one sense in which I think the idea of (but not actual) targets can be useful and that is in setting the scope.

If we want to find an improvement that is immense (versus small continual improvement) that can set the expectation of how we approach improvement, including an understanding that we are going to have to really make big changes in how things are done.

I have written more about this, here:

Deming on Targets

Basically I don't see that scoping "target" as really a target but it is similar so if you want to see it that way, then in that sense I can see a "target" as useful.

Related: Innovation at Toyota - Targets Distorting the System - Righter Incentivization

Saturday, April 04, 2015

The Value of Putting Pen to Paper

Comments on Learning by Writing… by Hand
The psychology behind the learning advantage of handwriting is starting to be understood... [Carol Holstead] "It turned out my theory was right and now is supported by research. A study published last year in Psychological Science showed that students who write out notes longhand remember conceptual information better than those who take notes on a computer."
I am also a fan of technology. And also a fan of learning and paying attention to research. Pen on paper has advantages for learning that technology has yet to equal. At the same time technology has many advantages also.

We seem to understand the advantages of using technology fairly well but under-appreciate the advantages of pen on paper. To make sure we don't lose out due to this bias we should think before we accept that pen on paper isn't worthwhile.

From a post I wrote in 2005, Measurement and Data Collection

I believe, it is better to focus on less data, really focus on it. My father, Bill Hunter, and Brain Joiner, believed in the value of actually plotting the data yourself by hand. In this day and age that is almost never done (especially in an office environment). I think doing so does add value. For one thing, it makes you select the vital few important measures to your job.
Lots of data will be kept in computers and that makes sense. But putting pen to paper has value that we too quickly dismiss.

Related: Experience Teaches Nothing Without Theory - The Illusion of Knowledge - Write it Down to Improve Learning (Ackoff)

Monday, December 01, 2014

Data Must be Understood to Intelligently Use Evidence Based Thinking

All metrics are wrong, but some are useful
Metrics might tell you something about the world in a quantified way, but for the how and why we need models and theories … metrics are generated must be open and transparent to make gaming of the system more difficult, and to expose the biases that are inherent in humanly created data
True, understanding the proxy nature of data (and how well or questionably the proxy fits) is important.

Data can't lie but we often make it easy for others to mislead us when we don't understand (or question) what the data really means (what operational definitions were used in the collection, etc.).

Related: Operational Definitions and Data Collection - Actionable Metrics

Friday, February 28, 2014

The damage caused by "Management" by targets is much larger in dysfunctional organizations

The damage caused by "Management" by targets is much larger in dysfunctional organizations - they are also more likely to be given more importance by dysfunctional organizations, that is a bad combination. In a great organization with an strong understanding of systems, respect for people, no pay based on "performance," an understanding of data and variation... then damage managing by targets does is much smaller. But the number of those organizations is not huge.

Reaction to: Target Setting, Cause and Effect

Related: Setting Goals Can Easily Backfire - I achieved my goal by not my aim - Be Careful What You Measure - Targets Distorting the System

Tuesday, June 11, 2013

Providing background material in advance of discussions

One place I see for improvement is an actual underuse of email. In meetings when lots of data is provided on some issue, that normally would be better handled by an email in advance (or could be some reporting system or whatever - but an email to look at urls of certain data...) to let people review the data.

Also preceding an urgent face-to-face, cell phone... with background info is often more helpful than trying to talk about stuff that is best digested by someone sitting and thinking. When on the phone or in the presence of others we often have a tendency to need to feel the space with noise/talking and deep thought is not likely, unless you already have deep thoughts on the topic and are just thinking of how to tweak those based on comments.

I see a systemic failure to provide background material in advance of discussions in many organizations.

Posted in response to How can you bring standard work to communication [removed broken link]?

Related: Process Thinking: Process Email Addresses - Better Meetings - Effective Communication is Explicit

Monday, April 08, 2013

Remembering George E.P. Box

George Box passed away last week and a long (1919 - 2013), rewarding and productive life. His obituary ends with "a last message from George" (quoting Cole Porter's song - Experiment).
“Experiment! Make it your motto day and night. Experiment, And it will lead you to the light …Be Curious, …Get Furious… Experiment, And you’ll see!”
The full text of the song is quoted in Statistics for Experimenters (a book by George, my father and Stu Hunter on using design of experiments to improve). The song is included in the De-Lovely soundtrack.

If you want to honor the memory of George, contributions could be made to

  UW Foundation - George Box Endowment Fund (link to donate - include George Box Endowment Fund in the box for instructions) US Bank Lock Box 78807, Milwaukee, WI 53278.  This fund was started some years ago with the intention of assisting graduate students. It is a permanent endowment fund, so contributions to the fund are added to the principal and the annual earnings of the fund are used to support the fund purpose. The purpose of the fund is to support activities of the Statistics Department with a primary (but not exclusive) focus on activities of direct benefit to graduate students.  Recipients will be selected by the Department faculty (or their designates) with input from Departmental graduate students."

  Agrace HospiceCare (link for donating online), 5395 E. Cheryl Parkway Madison, WI 53711.

Wednesday, July 11, 2012

Leadership: Taking Action When Others Are Unsure


When things go bad
Plan for the things you know can happen. Weather emergencies, hazmat spills, and fires come to mind. Rick Rescorla's plan for what Morgan Stanley workers should do if the World Trade Center was attacked, is one reason why that firm only lost six people out of almost three thousand on September 11, 2001. 
Include some thinking about how you will know what you're facing. While the Port Authority officials were broadcasting "stay put" messages, Rick decided that an attack was happening. He activated his plan, marching his company employees down the stairs, two by two.
I can't recall now, but I think there are studies that show people will be lulled into not acting when a group of people is around (who are also not taking action).  So something is introduced that if they were alone they would react (say leave the trade center, or investigate a seemingly risky piece of data).  But if there is a group people become more passive - thinking the non-action by others means they are over-reacting.  So they then don't react (thus re-inforcing everyone else decision not to act).

This is one of those times leadership really matters: someone not afraid to take action and potentially criticized for going against the consensus group decision to not act.

If immediately upon suggesting decisive action, people jump to support the idea it is likely they all would have done so alone but were intimidated by the non-acting group.  Either that or they respect the leader and decide to support them while not sure it is really needed.

Related: Leadership is the act of making others effective in achieving an aim - Leadership Leverage Points - quotes on leadership

Sunday, March 11, 2012

The Potential Benefits, Risks and Folly of Stretch Goals

Some excepts from, The Folly of "Stretch Goals", visit the link to see the full discussion:
Jon Miller: Stretch goals are fine, but gaming the system, sandbagging, achieving the stretch goals through heroic effort, etc. are bad because this is not sustainable. In terms of excessive risk taking, this is a question of the risk-reward calculus and the person’s degree of risk aversion. It doesn’t take a stretch goal to make Enron leaders cheat when their auditors are turning a blind eye. They stole because they could, not because a leader set stretch goals for them. If the governance around the goals are solid and the downside of risk are significant, people will pursue stretch goals in a way that is not destructive. ... 
Dan Markovitz: However, if you had the opportunity to make a HUGE bonus — millions or tens of millions of dollars — for achieving certain stretch sales targets in China, for example, you might be sorely tempted to act differently... 
Jon Miller: But my point was that cheating is not caused by stretch goals, it is caused by poor governance around the performance and rewards process... The more interesting question is why leaders continue to set up such systems. Are they stupid? Evil? Or do such systems produce results?
I think stretch goals are fine when people understand - they are giving scope to the effort. If I want breakthrough improvement quickly it may mean considering radical solutions. That can be helpful to shape people's vision. But there are risks. As Brian Joiner said there are 3 ways to improve figures ("results")
Improving the system is far more difficult than the first 2. Cheating can be encouraged by managers. Stretch goals can increase this encouragement. A culture that pushes the right values and discourages the wrong ones can discourage cheating. Understanding variation is very helpful (it both dramatical reduces silly reaction to variation - the fear of those silly reactions often cause people to cheat "distort the figures or system" ).

An understanding data is only a proxy for the real situation (the number is not real situation) is helpful as is understanding the arbitrary goal is essentially meaningless (it exists to give scope to efforts not to be met - 67.3% improvement when 75% improvement was the "goal" is not failure - an understanding of variation would assure this mistake was not made).

The problem is many organizations are ruled by spreadsheet managers that don't understand variation, are ruled by the tyranny of arbitrary targets (bonus and promotions)... In these situations goals do often become a big part of the reason for cheating. Stretch goals can help shape the effort. The risk (and much more common result, I think) is that they result in distortions of the system and data to achieve those results.

To answer Jon's question I think you can use goals and incentives to reach numerical targets. The risk, as Gipsie Ranny says is the organization may be ruined in the long term. But if the executives are fearful and have large enough incentives to achieve numerical targets they goals and targets can achieve the goal - but at a great cost, I believe. I believe they are more ignorant than evil (though some know the damage they are risking or causing).

A strong management system reduces much of the potential negative consequences of targets. A big problem is those organizations that most rely on targets are those that are least protected from the risks of using them.

Related: The Defect Black Market - Targets Distorting the System - The Problem with Targets

Thursday, October 06, 2011

Lying with Statistics


Response to: Great example of "Lying with Statistics"

My view is closer to Rip's.  Deceiving people is not alleviated by being "truthful" but misleading.  As with many things where you draw the border is often challenging.  I do like putting the claims of lying on a person - not on data.  Data can be wrong.  It can't lie.  People can lie.  People can also mislead.  And very often people can be mislead (by those intending to mislead them and those that failed to understand the data in the first place and then used the data in a faulty way to support their mistaken notion).

Those of us reading the messages in this group (statistics group on LinkedIn) are not likely to fall into the being mislead camp often.  But my experience is that is by far the biggest problem.  People not having numeracy and being mislead all the time do to their lack of understanding (either intentionally, or through ignorance of theirs [or the person presenting the info to them]).

Related:  Bigger Impact: 15 to 18 mpg or 50 to 100 mpg? - Understanding Data -

Preaching False Ideas to Men Known to be Idiots

Friday, August 05, 2011

Experimenting to Discover

Causal Reasoning in Science: Don’t Dismiss Correlations (the broken link was removed)
Box, Hunter, and Hunter were/are theorists, in the sense that they don’t do experiments (or even collect data) themselves.
...
Science is about increasing certainty — about learning. You can learn from any observation, as distasteful as that may be to evidence snobs. By saying that experiments are “necessary” to find out something, Box et al. said the opposite of you can learn from any observation.
William Hunter was my father. He did many experiments. George Box did many experiments. You are entitled to your opinions obviously but the claim that they only dealt with other people's data is not accurate. It is true they were world renowned experts on experimenting and had many people consult them about their experiments, for help: designing them, analyzing them, what to do next, how to improve the process of experimentation in their organization, etc.. While it seems to be implied in the post that such consultation was a reason to distrust their thoughts on experimentation I hardly think that is a sensible conclusion to draw. Most of those they helped were running experiments in industry, to improve results (not to publish papers).

They were, and are, applied statisticians (and though I am obviously biased, I think many would agree, 2 of the most accomplished in that field in the 20th century). What experiments need to be done is critical for an applied statistician. What matters is making improvement in real world processes. If you don't run the right experiments, you won't learn things to help you improve.

They worked on the problem of where to focus, in order to learn, quite a bit. One significant part of there belief was to have those involved in the work do the thinking about what needed to be improved. This isn't tremendously radical today but in the past you had many people that thought "workers" should do what the college graduates in their office at headquarters tell them to do. Here is one of many such example, from Managing Our Way to Economic Success by William Hunter:

The key is that employees at all levels must have appropriate technical tools so that they can do the following things:

- recognize when a problem has arisen or an opportunity for improvement exists,
- collect relevant data,
- analyze the situation,
- determine whose responsibility it is to take further action,
- solve the problem or refer it to someone more appropriate...


I don't have the book in front of me, but doesn't it start with an example on learning where you can use inductive reasoning and from the facts that you see you can draw conclusions and construct a theory that fits the facts. If so, it seems to call into question the idea that they claimed "[the] opposite of you can learn from any observation." is not actually accurate. They understood you can use inductive reasoning to create theories. You then use experiments to test theories.

The books is called Statistics for Experimenters, right? Not statistics for drawing conclusions when not doing experiments. When you are experimenting you can test whether beliefs you have are accurate and you can learn about things you try. Smart people can make guesses what will happen and be right. I know the authors would believe those knowledgable about the system in question are well suited to determine what variables to test. It is that knowledge that will lead to experiments that are likely to be effective.

The authors of the book were trying to help those that often failed to learn as much from experiments as they could. Far too many people still don't use the most effective statistical tools when experimenting.

They emphasized, consistently, the need for those doing the work to involved in the experiments. The job of statisticians was to help in the cases where advanced statistical tools and knowledge would be useful. The reason for those who do the work (are familiar with the process) is because they have knowledge to bring to what should be tried in experiments.

When I read through The Scientific Context of Quality Improvement, 1987 by George Box and Soren Bisgaard it seems to me it discusses the types of issues you raise: how do we learn without experimenting? I am not sure if it is just me, or if it clearly addresses that issue. Here is another, Statistics as a Catalyst to Learning by Scientific Method by George E. P. Box. And another, Statistics for Discovery.

There are many other sources, I am sure. They understood the importance of learning as much as you could from available sources. They just also understood the importance of experiments and learning the most you could from experiments. And the book, Statistics for Experimenters, was focused on the most effective ways to improve using statistics to learn from experiments..

Here is what Box, said in his own words about the objective (and it isn't proving the hypothesis):

[too many people ]"can’t really get the fact that it’s not about proving a theorem, it’s about being curious about things. There aren’t enough people who will apply [DOE] as a way of finding things out"


Statistics for Experimenters: Design, Innovation, and Discovery shows that the goal of design of experiments is to learn and refine your experiment based on the knowledge you gain and experiment again. It is a process of discovery. That discovery is useful when it allows you to make improvement in real world outcomes. That is the objective.

Saturday, April 09, 2011

Discrimination and Data

A Discriminatory Conundrum

The total American workforce has remained relatively constant over the last ten years – roughly 131M employees. The number of EEOC claims over the last ten years has increased by roughly 25% to almost 100K a year.


I would think these suits increase when there is a bad employment marketplace, but I don't have any data on that it just seems logical. When times are good all sorts of people get good jobs. When times get bad, jobs are hard to find, promotions are hard to find, people get demoted, bosses have to cut budgets and people. People are also very stressed out and it isn't surprising to me that lots of bad outcomes come from that including people feeling slighted, people getting into escalating cycles of bad words and actions, and more lawsuits.

In addition to the macro-effects of the economy data has quite a bit of variation. Often lawsuits over class differences get data that shows difference in result and claims this shows discrimination. While if we showed the same data for the first letter of the grade school the person attended, or eye color, or the state the person's Mother was born in we would see lots of variation in data that people would see as worthy of payment for the discriminated against class (though in those cases no-one would actually believe it). xkcd took a comic look at data analysis recently.

I am skeptical of short term variation having much meaning with this data. I would want to see it charted over time and then analyzed for having an indication it wasn't just random variation of a stable system.

If the data doesn't indication a special cause (for the variation in the data) that tells you that special cause problem solving is not the best way to improve. In that case you still want to improve but you improve using common cause problem solving techniques. This particular dataset would undoubtedly (at least in my opinion) benefit from stratifying the data to identify segments where it is more of an issue. My guess is that there would be wide variation but the data analysis would show if this is true or not.

Thursday, August 19, 2010

SPC - Charting and Improving Results

Everett Clinic Video, Redux – The Need for SPC Thinking

Looking at 5.x% and comparing it against an arbitrary goal does little to tell us about the health of the work system. Is 5.x% the typical average performance? Is that much higher than usual?

This is a great opportunity to use the methods of Statistical Process Control. The main management decision is to decide "react" or "not react" to that daily data point. SPC helps us with this (again, Wheeler’s brilliant little book explains this far better than I can in a blog post).

If we choose “not react” because 5.x% is lower than the goal, we might be missing an opportunity for process improvement. Generally, it’s better to present more than one data point – even if you don’t do full-blown SPC, you should present a run chart.
Well put. A simple run chart can be very helpful. One of the uses is to identify special causes. And then to use special cause thinking in those cases. What is important about special cause thinking? That you want to identify what is special about the data point (instead of focusing on all the results as you normally would). What is important about doing that? You want to do it right away (not a week or a month later). Keeping the chart lets you identify when to use special cause thinking and react quickly (to fix problems or capture good special causes to try and replicate them).

You have to be careful as we tend to examine most everything as a special cause, when most likely it is just the expected result of the system (with normal variation in the data). Special cause thinking is not an effective strategy for common cause results.

Related: Quality, SPC and Your Career - Statistical Engineering Links Statistical Thinking, Methods and Tools