Lean

Why Do We Need to Set a Hypothesis?

Avatar photo By Jon Miller Updated on August 5th, 2018

Why do we need to set a hypothesis? Because that’s how the scientific method works. Why do we need to follow the scientific method? Because that’s how we solve complex or novel problems most effectively. Why is that true? Because we test and refine our hypotheses. We were going in circles. My customer was challenging the assertion that it was necessary to explicitly state that her organization’s method for problem solving be scientific. Resistance to using the term “lean”, Japanese terminology and common acronyms such as DMAIC or SMED is still fairly common. This objection to “scientific method” was a new one.

She argued that for her colleagues who were scientists, engineers or researchers with a background in the scientific method, emphasizing that phrase would seem redundant, too familiar, making the building of a new continuous improvement culture hard to sell. To workers without that education, the term scientific method would seem intimidating. The creative, artist types in her organization to chafe at the implied rigidity of “science”. Leadership would mentally position “science” below “strategic” and lose interest. Why not just specify the broad activities at each step, such as defining the problem in the customer’s language, clarifying the problem in terms of a gap, breaking down the problem into its components, looking for root causes, etc.? Her concerns seemed valid, but I suspected that solving the problem of resistance to the words “scientific method” would create other problems.

Without the discipline of the scientific method, we rush into solving problems. We rely on our experience or expertise. We think we know the answer or that we are smart enough to figure it out. IF we are not careful, we come to associate the solution to our own sense of self-worth or ego. Being right can even become more important than solving the problem. We may be right. Our intuition, creativity and experience may solve the problem. But we are all wrong at least half of the time.

What happens when we solve problems without explicitly following the scientific method?

  • We are wrong but the reason why is not clear
  • We need to justify, defend, make excuses or deflect blame for being wrong
  • We spend mental energy re-inflating our ego, recovering from being wrong
  • We become deaf to observations and insights of others

It’s OK to be wrong. It’s not OK to be wrong in the wrong way. There is a right way and a wrong way to be wrong. A hypothesis is a tool to help you be wrong in the right way. A hypothesis is a tentative, testable answer to a scientific question. It is an educated guess.

What does it mean to “be wrong in the right way”?

  • We propose a tentative, testable answer to a question
  • We know we could be wrong and are not threatened by this
  • We intend to test it as being right or wrong, not defend our belief
  • We don’t need to make excuses when we are wrong
  • We are clear about what cause-and-effect relationship we are testing
  • We gather and share data in a way that invites others to examine and contribute

The word hypothesis comes from the Greek word “foundation”. A hypothesis is not meant to be the final answer, only the beginning of investigation. The act of setting a hypothesis allows us to structure our thinking in such a way that we can safely learn the right lessons from failures. This seems like a small price to pay for making and effort to get over any aversion we might have to adding “scientific method” to our permanent vocabulary.


  1. Tracy Defoe

    August 7, 2018 - 1:15 pm
    Reply

    I have coached research scientists in the Kata methods to improve their communication and management skills, and one thing they always bring up is that in their way of thinking, the hypothesis is something to be disproven. That is, you run experiments in an effort to disprove your hypothesis. So we usually settle on framing the Improvement Kata as a practical application of scientific methods, and we can usually all work with that.

    Good article! I had a learner once object to the term “challenge” because that was their corporate term for “problem with a client.” There were no problems, just challenges. She really got stuck right at the top of the board with the Challenge. We settled on Big Goal or something like that to make it work in their culture. You have to listen.

    • Avatar photo

      Jon Miller

      August 8, 2018 - 11:18 am

      Hah! I had never heard the one about “challenge” but I can see their point. Words are important. They have meaning, and elicit thoughts and emotions.

      Great point about the hypothesis being something that we aim to disprove when doing science. It’s a lot easier to accept being wrong when that is explicitly a possibility that you are testing for.

  2. France Bergeron

    August 8, 2018 - 5:18 am
    Reply

    There is a lot of confusion about science. There is no “scientific method”. What is called “scientific method” is “experimentation” which is only one tool (one method) used in science to disprove your hypothesis. Lean is the science of work, here’s how:

    To be called a science, you must have the following 3 elements:

    1. Understand something scientifically (remove assumptions, it’s not WYSIWYG and have an explicit theoretical context). In Lean, work is understood as flow (not waste, not cost, not quality) and the theoretical context is optimization of the flow of value to client with equals who collaborate (Lean is collaborative science). You reduce waste, reduce cost and increase quality so you flow more value to your client.

    2. Use rigorous methods (or tools): this is where tools (methods) such as experimentation, observation (ethnography) at the Gemba, visuals are used. Without #1 (understand what you’re looking at (flow)), your experimentation will go in all directions and you won’t have repeatable results.

    3. Share scientific explanation. In Lean using A3, case studies, videos for example.

    To learn more about Lean as science, see our paper “Lean: it’s not rocket science, it’s work science” https://bit.ly/2M1AhrK

  3. Avatar photo

    Jon Miller

    August 8, 2018 - 11:55 am
    Reply

    Thanks for your comment France.

    I think the global community of scientists would agree that for the past 300 years or so there has been a consensus method or approach for getting to closer truth through observation, measurement, experimentation, and the formulation, testing, and modification of hypotheses. Is there no such thing as the scientific method?

    I agree that Lean is the result of approaching work, or the fulfillment of any human objectives, scientifically.

    The paper you reference is curious. It seems to create a dichotomy where none is needed. Lean and science both examine the situation in context of what we are trying to achieve, the current situation, the reality, the facts, to look for truth.

    Neither Lean nor the scientific method ever attempt to find “the perfect truth” so point in the article seems like a straw man argument.

    The process of finding truth (or knowledge if you prefer) involves people. Lean does recognize that the people doing the work bring expertise. Sometimes we arrive at truth through social consensus. Sometimes because facts are exposed through the rigor of experts.

    It is idealistic but not based in reality or supported by facts to claim that “Knowledge in Lean is socially constructed by all as equals and therefore cannot be framed as ‘truth’.”

  4. Brian M. Kennedy

    August 10, 2018 - 9:55 am
    Reply

    Hi Jon, I am going to have to side with your customer on this one. While I agree with you on the important attributes of a good problem-solving process, you have not made adequate case that the phrases “scientific method” or “hypothesis” need to be involved. Not only for the reasons your customer gave, but for another that will likely come out in the other comments here: people have different (mis)conceptions about the scientific method that can work against the good problem-solving practices you’re trying to instill in people. As evidence, we have published a comprehensive book on collaborative problem-solving that teaches all the good practices you mention (and more) and never uses either phrase.

    Further, let me argue that ONLY testing hypotheses for failure is actually an unnecessary limitation that will slow your teams down. Yes, there are many reasons why only doing that is better than only testing for success… however, having both tools in your toolbox is best. In the end, your decision space is made up of three things: the sets of things you know will work, the sets of things you know will NOT work, and then all the rest that you don’t really KNOW whether it will work or not. And our teams generally have very limited time to solve problems (or make decisions)… so, as quickly as possible, we need them to focus in on that “don’t KNOW” portion and gain clarity by EITHER identifying more stuff that they know will NOT work OR more stuff that they know will work. The goal is to make sure they find some sets of things they know WILL work and that will satisfy the customers/stakeholders (“Success is Assured”), such that they don’t end up choosing stuff they don’t KNOW hoping it will work (“Wishful Thinking”). In many cases, testing for failure will eliminate more of that “don’t KNOW” stuff faster; but there are plenty of cases where testing a REGION (not a POINT) for reliable success can eliminate more “don’t KNOW” stuff faster. The key thing is to stop thinking in terms of testing POINTS for success… instead test SETS (regions of the decision space), either for success OR failure.

    I will argue the primary advantage of encouraging testing for failure is that doing so tends to move people to testing sets, whereas testing for success tends to move people to testing points. But the issue is “sets” vs. “points”, not “for failure” vs. “for success”.

  5. Avatar photo

    Jon Miller

    August 10, 2018 - 11:25 am
    Reply

    Hello Brian

    Thanks for your comment.

    Your argument seems to be that as long it is scientific thinking, we do not need to call it scientific thinking. My argument is that avoiding the s-word because some people don’t like it is just avoiding the problem.

    I agree with you on the importance of set-based approach.

Have something to say?

Leave your comment and let's talk!

Start your Lean & Six Sigma training today.