Problem Analysis

Quick reference

• 00:05 Hi, I am Ray Sheen.
• 00:06 Well, it's finally time to enter into the Analyze stage of our Lean Six Sigma
• 00:11 project.
• 00:12 Now to do this, let's start by clarifying
• 00:15 some aspects of the problem solving process that we are using.
• 00:20 The Lean Six Sigma methodology is a problem-solving methodology.
• 00:24 If there's no problem, well then there's no reason to do the project.
• 00:28 Lean Six Sigma does not follow the lucky guess approach to solving a problem.
• 00:33 Rather, relies on understanding the details concerning the problem in order to
• 00:38 determine what type of solution will work To actually solve the problem.
• 00:42 This type of problem solving method requires both inductive and
• 00:45 deductive reasoning.
• 00:47 We use inductive reasoning to take a specific observation and
• 00:50 generalize that into a theory of the problem.
• 00:53 And we use deductive reasoning to take a general theory and
• 00:57 test it with specific results and observations.
• 01:00 Most problem-solving methods used in business today,
• 01:04 and that includes the Lean Six Sigma methodology, rely on a combination of both
• 01:08 operational systems analysis and mathematical data analysis.
• 01:12 That is why Lean Six Sigma uses both the value stream map to understand
• 01:17 the system and collects data for statistical analysis.
• 01:20 While we intend to be methodical and logical, we can still apply experience and
• 01:25 judgment to guide our analysis.
• 01:26 That is often what helps us to know what questions to ask.
• 01:30 In fact, many of the business problem-solving methodologies rely on
• 01:35 the scientific method of analysis.
• 01:37 This provides a logical approach for organizing our problem solving process.
• 01:42 The scientific method is defined as a method of analysis that involves
• 01:46 systematic observation, measurement, and experimentation.
• 01:50 It includes the formulation, testing, and modification of hypothesis.
• 01:55 As you can see, the process starts by asking a question and
• 01:58 then creating a hypotheses for the answer to that question.
• 02:02 Next, develop an experiment to test the hypothesis,
• 02:05 make certain the experiment is valid, and if not, develop a new experiment.
• 02:09 Then, use the experiment to collect data and analyze the data.
• 02:12 From the analysis, draw a conclusion about the validity of the hypothesis,
• 02:17 or go back and create a new hypothesis.
• 02:20 You can see this combines both inductive and deductive reasoning.
• 02:23 Creating the hypothesis is inductive and performing the tests is deductive.
• 02:28 The approach also organizes the problem-solving activities by giving us
• 02:32 a plan of action.
• 02:33 We know what to do next, and the methodology is a proven and
• 02:37 well established approach.
• 02:38 So it gives us confidence that it will lead to a solution.
• 02:42 Let's dig a little deeper into the concept of questions and hypotheses.
• 02:47 Problem analysis is an inquiry process.
• 02:49 A problem is a difficulty, and we need to analyze the problem when we are not
• 02:53 certain of the best way to remove the difficulty.
• 02:56 Once we know what to do, we no longer need to analyze the problem,
• 03:00 we just need to do the work to fix it.
• 03:02 So as part of this inquiry process, we first need to formulate the question.
• 03:07 Knowing the question helps us to know the goal or purpose of the experiments and
• 03:12 the data analysis.
• 03:13 We don't run the numbers just to practice doing math,
• 03:16 we're trying to answer a specific question.
• 03:18 I've found that using the cause an effect format helps to formulate the question,
• 03:24 and by that I mean does this question cause this effect?
• 03:28 We can then restate the question into a set of hypotheses.
• 03:32 These hypotheses look at both sides of the question.
• 03:35 We call the two perspectives the null hypothesis and the alternate hypothesis.
• 03:40 The null hypothesis is always framed in such a way as to say that the dependent
• 03:44 variable is not affecting the independent variable.
• 03:48 In other words, there is nothing to see here.
• 03:51 And of course, the alternate hypothesis is the opposite.
• 03:54 It states that the dependent variable is affected by the independent variable.
• 03:59 We have an entire course that explains hypothesis testing,
• 04:02 not just the writing of the hypothesis, but how to prove or
• 04:05 disprove the hypothesis with statistical data.
• 04:08 But let me give you a quick overview.
• 04:10 Statistical hypothesis testing allows us to test a claim or hypothesis about
• 04:16 a characteristic in a data set and determine its statistical validity.
• 04:20 Based upon the hypothesis test, we can accept or reject this claim.
• 04:25 And while we do this with a sample of the data,
• 04:28 we can draw conclusions about the entire population that the sample represents.
• 04:33 With that analysis, we can answer questions posed by our hypothesis and
• 04:37 determine if there really is a difference between this and that.
• 04:41 Knowing what the difference makers are will normally lead us to the cause or
• 04:45 source of the problem.
• 04:47 Now, we normally use hypothesis testing to confirm our suspicions,
• 04:51 not as a tool to discover new information.
• 04:54 That is why we need to apply our judgment and
• 04:56 experience to a situation to create a question on hypothesis.
• 05:00 When doing a Six Sigma hypothesis testing analysis,
• 05:03 I recommend that you intentionally create several hypothesis and then use
• 05:08 the data that was already collected in the measure phase to test these hypothesis.
• 05:13 Don't worry, In the hypothesis testing course,
• 05:15 we will discuss the various analysis.
• 05:17 If you already have your data set,
• 05:20 each hypothesis test will only take a few minutes to statistically analyze the data.
• 05:25 And that is what makes this approach so powerful.
• 05:28 The data tells us what is relevant.
• 05:30 It's not based upon a lucky guess or a forceful PowerPoint presentation.
• 05:35 Rather, the data tells us what is happening and
• 05:39 leads us to the cause or causes of the problem.
• 05:42 In this Analyze stage, we want to discover the underlying cause so
• 05:46 that we fix the correct thing once we get to the Improved stage.
• 05:50 The scientific method is a proven problem-solving approach that will lead
• 05:55 us through this stage.

Lesson notes are only available for subscribers.