# 9.7 Hypothesis Development

Video Activity

Join over 3 million cybersecurity professionals advancing their career

Sign up with

Required fields are marked with an *

or

Already have an account? Sign In »

Video Transcription

00:00

Hi, guys. Welcome to analyze phase hypothesis development on Katherine MacGyver and today you're gonna understand hypothesis, development and have the ability to documents. Ah, hypothesis.

00:11

So getting started. What is a hypothesis? My offices is most simply put is a question. It's the idea of a relationship between independent and dependent variables. So we think that because of X, why occurs if you remember back to our y equals f of X?

00:30

This is where it started being articulated as multiple variables.

00:35

So hypotheses do not have to be a wonder. One relationship. They can be a mini to one. So we think because of

00:42

the weather and the cash register and the customer base, this is why we have customer complaints so

00:50

can beam Anita. One can also be one too many. So you have independent variables that Dr multiple dependent variables. That's a little bit harder to tease out when you're doing your hypothesis development. But ah, hypothesis is a proposed explanation based off of limited information. So when we think about proposed explanation,

01:08

if we know this solution, we don't need a hypothesis development

01:12

so much like our just to do it projects in our quick hits. If you know how to fix the problem, then just fix it. If you know what is causing this or what independent variable is driving which dependent variable, then just deal with it. Um,

01:30

another thing that is important to keep in mind when we're talking about hypothesis development is that it doesn't account for correlation and causation, at least at this phase. Once you start getting deeper into some of the more sophisticated statistical test, you may be able to tease out Corliss and correlation and causation.

01:46

And correlation is when you have variables that move in conjunction with each other, they can either be in peril or they could be paradoxical where they move opposite each other. So if you see an increase in X, you see a decrease in why or if you see an increase in X,

02:01

you see an increase in white. So correlation moves together frequently. Looks like causation because of the patterning and then causation is something causes something. So you have a predecessor. That is what triggers this activity. So if I change X, why does this

02:21

causation is what we're looking for and lean and six Sigma.

02:24

We're not gonna drill into too heavy and it in yellow Belt. I did promise you guys at the beginning that we're not gonna do a lot of statistics, but it is important to understand the idea. So when we're developing hypotheses, we're developing questions about the relationship between our variables. Do we have it have lengthened

02:44

cycle time because we had unscheduled downtimes?

02:47

That's a pretty straightforward we know the solution for, but maybe we don't.

02:53

So when we're talking about why do we want to develop hypotheses? We're now getting more into that statistical room. So hypothesis Development tells us, tells us if there is a statistically significant difference between two datasets. So when we talk about statistically significant, what we're talking about is

03:12

everybody has good days.

03:14

Your organization is going to do better. On some days, as compared to others, we go back to process entitlement. So did the changes that we do actually manifest mathematically as a difference between the two datasets. So, um,

03:30

translated, it tells us whether or not our changes had an impact. So if we change acts or a variety of exes, do we really change why? Or is this noise? So those idea of noises that you always have process, very ability.

03:46

So was it. We changed X, and then the next day we had a really great why

03:52

could it have been attributed to something else? That's what hypothesis testing tells us is it gives us a confidence interval of whether or not the work that we did actually caused the results that we're seeing

04:04

before we can test a hypothesis. We have to draft a hypothesis, so you will always have a minimum of two hypotheses. Um, I have seen upwards of 12 in a study, but two is too is good for starting. Maybe you want to get to three, you're going to want tohave. Um,

04:24

hypothesis, more hypotheses if you have more variables and you have a couple of different ways of wording it. So, um, your first hypotheses you will always be your h not. Or you're no hypothesis. This is that nothing changes. This hypothesis states that the work that we did did nothing.

04:42

Um then your alternative hypotheses or your age sub one or h sub.

04:46

However many is that question of the relationship between X and Y. If we d'oh x, why will do this s O. There are a couple of qualifiers when we're talking about alternative hypotheses in how we dictate the type of test that we do as faras.

05:05

If we do this,

05:08

it will improve. If we do this, it will decrease. If we do this, there will be a change. For the sake of this conversation, we're going to say two tailed, which is simply if we do this, there is some change neither positive, more negative.

05:25

Um, when you're talking about how it actually looks a tch not or the null hypothesis for this example is adding an additional shift to a production line will not change production. So what we're saying is, the activity that we're doing will have no impact

05:45

statistically on our production rate. So conversely, are our alternative hypothesis is adding an additional shift. A production line will change production, so please, no here

05:57

that we don't say will increase production or will decrease cycle time. We just say that there's going to be a difference between the two datasets.

06:05

So

06:06

a quick break. Let's test your knowledge. Which of the following statements are hypotheses? So we have one. I think my apartment has a mouse because my roommate is a slob to the customers are unhappy with the quality of pizza. And three, my community center needs a rec center and library.

06:28

So looking at one, this is a hypothesis. So you have a dependent and an independent variable. Mice are going to be your dependent variable. Being a slob is going to be your independent variable. So if we were to develop a possible hypothesis statement from this, we would say because of my roommates cleanliness.

06:47

We have mice, which means that if we change

06:50

your cleanliness, I'm either more or less you will have more or less mice. Arguably number two. The customers are unhappy with the quality of pizza. This is actually also Ah hypothesis statement. So you have your dependent variable, which is unhappy, and the quality of pizza is your independent variable.

07:10

The assumption here is if you improve the quality of pizza,

07:13

you will improve your customers satisfaction. So this is again a one tailed. We have a relationship between there, but we could say improve quality of pizza will change customer satisfaction on, and then the last one. My community needs a rec center in library. This is not a hypothesis

07:30

because there is no assumption of relationship between the two. This is simply a declarative statement.

07:39

So when we're talking about hypotheses, remember, we're using this for relationships between variables you want dependent and independent variables. Where we get those variables from can be our root cause analysis. It can also be in our measure phase. But when we started talking about what do we think the drivers of these processes are?

07:58

That's what these variables are.

08:00

And then once we develop our hypotheses, we're going to use that and the Associated Data Collection to test for statistical significance. So that's hypothesis development. Our next module is actually going to be on hissed a grams and defect prioritization, so I will see you guys there.

Up Next

Similar Content