Statistical sampling This is very important when you are doing inspection It is very important when you are doing control quality It is very important when you are doing delivery validation of scope. These are choosing a part of a population of interest for inspection the sample is taken to measure control and verify the qualities when you have a batch of thousand pills or a batch of thousand years a batch of 1500 cards you cannot you will definitely have a quality check for all of it independently, but you would have some sampling to see that how the process is you when you are running the report and the process purchase order daily you cannot go for all 5000 purchase order in terms of sampling. What you would be doing is taking the in terms of the population what you would be doing you will be taking some sampling out And then based on the sampling you will see whether the person or the people or the group of people or the team, are they following it?

Are they following the results following the identified process or not, is the deliverable is as the quality that we have agreed for sampling is the selection of subsets of individual from within the statistical population to estimate characteristics of the whole population. So, I would say that, if this badge is 4000 and I have selected 190% have passed 5% are okay and 5% are bad. So, I would say as part of the whole bad, almost 50 are will be failed in terms of their final quality deliverable. Determine the variables to major from the beginning. We have to move away from qualitative terms terminology, we have to move away from my My word against your word we have to move away from that what is my understanding of God and what is your understanding of God we need to put figures who are clear measurable figures to determine the variable and to major determine the population describes the method of sampling, select the size collect and get the data from the sample.

It is largely used for quality. The problem is what is good coverage. I have seen data in the last election where the data is for 10,000 for one or sorry 1500 for a constituency where you have 400,000 people, no you cannot have it. Then I see a data that it is for a specific area a specific kind of people in this area when you are doing a constitute sisters Pampling. This area is not happy with you So it is very interesting when this sampling comes and plays a very interesting role in our lives in specifically when we see it and that's if you find multiple TV anchors, they're standing in the same area in the same cost agency and the results are completely different The problem is their coverage the sample that they are selecting is not good enough it is biased towards anything. So, they have gone into an area which is not very happy, they have they have gone into area which is ethnically different, the results will be different.

So your sampling has to be very good. It's a very, it's a challenge to have a good sample in terms of expecting because you are whatever the results you will get from that sample, you will assume it for the whole group. So if your sampling is not it does not have a nfx coverage or a good coverage. Your research will be biased or at least you will have a wrong results in terms of the final data representation. Well, now it is time to go into a little bit deeper in terms of sampling. The idea is very simple that I want to give you the overview of how you will be selecting a sample or if anyone of your team who is looking after the quality control within your team me is part of inspection team of your delivery or quality control, how should they be doing it and if they are giving you any specific terminology, I think you would have a fair idea, but we have to kind of assembling one is probability sampling and other than non probability.

In terms of probability sampling, we usually have systematic sampling, random sampling, stratified sampling, and non probability sampling is convenience quota system. All these are nonprobability samples. The idea behind the sampling is to obviously ascertain the fact whatever we the results we will get from the sample we will assume or over the purple or all over the population. So, first of all we have a systematic sampling, starting randomly and select the next item systematically. So, you have a population of thousands. So, every 50 of you will select this will give you greatly unbiased and responsive representative I would say popular sampling because this these are the two important factors when you are talking about the sampling dish it should be unbiased and it should be representative.

The next one is random random select the net respondent next respondent from the population to maximize the representation and be unbiased So, you will be selecting any without any biasness or pre assumptions or any questionnaires or anything, this is like a lottery you will get a very fair number of representation. Then we have a stratified sampling or an ad, the advanced version is clusters to divide the population in sub population also known as strata and then select from each randomly. So, we have selected all males or females or a 25 or 230 stuff like that, and then we select randomly from them. It will also give you pretty representative and unbiased data and purpose if I would not say that this is again a nonprobability sampling but it would definitely fall into that reason being you have identified the predefined criteria and then individuals are matched over that that what will be the So, anybody who is aged 25 and above masters living in this area, so, these are a purposive sampling, then we have a convenient sampling you would not go You are not actively trying to do NFL you whatever are closed by you are collecting this information in this type of a sampling we have a self selection bias usually come in because if I find that questionnaire or these sampling topic is chocolate, I would like to jump in if I find the sampling topic is anything which is controversial, I might say no.

So, this is self selection bias comes in. So, these are the different kinds of way how you will actually gather the sample of the population. Then well while you are getting the answers you have to kind of the responses one is attribute fleet sampling where the answer is yes or no true or false pass or fail you will actually fail or you will say that this product passed the test or this product failed the test. The next one is variable and continuous sampling I'm sure we all have seen multiple time bad Good, very good. Satisfied very satisfied. Starting from unsatisfied very satisfied.

Unsatisfied satisfied, very sad but first a scale of one to five a scale of one to 10 and you try to degrade yourself or that sample that how good it is. So it is in time this is what usually said. When you have a space that anything above 80% is okay for you to say as it is for fulfilling the 80% of this. You usually do these kind of a continuous sampling when there is a grading is happening. When you do like a classroom test ABCD so some of his past and B some as Pastor see some as pastors d. So, in the past second, then you usually get this in training schedule, then you really get this in terms of courses. So, here the idea is that you are actually quantifying the degree of conformity based on the continuous scale over the period of time.

So, this is how this is about sampling as I said, we have a population and we cannot go back to the population population in any big part, we have to have a sample we want to make sure that our quality is up to the mark, we want to make sure that whatever inspection We are, we are doing it has a proper unbiased and representative sampling. So there are multiple ways of doing it. As I have explained you five and six of it, and then this is how you will get in there. This is where I close my topic for today. For sampling, we started from the sampling, how you will do the sampling and these are some of the kind of the sampling definitely This is absolutely not if you will into some sampling code that you would do in statistical statistical course, statistics course, there you will be having a detailed conversation of all this and why you should be using sesac systematic sampling Why would you be using random stratified there are be there are purposes for these things you have to use systematic you have to use random sampling system I like I can give you one quick example systematic is easy when you are actually doing fruits and vegetables and when you are doing actually the manufacturing.

Random sampling is good when you are doing population and you want to you are in part of any school call. You want to have some question answers. certified is a it's a time taken you need to first of all understand your population then you need to make a strategy out of it. So, there are specific reasons question answers, types of sampling, then you would be looking at when you are actually doing as part of your statistics. Here, as I said, project manager is a breadth of knowledge rather than a depth of knowledge, you should have a fair idea about everything that is happening on your project, you would have subject matter experts on your project that would be doing a quality control, sampling and inspection and everything. You what you need to know you need to understand the terminology you need to understand what they are doing.

And this is where I believe this topic. Whatever I have covered in this topic is more than enough for you to bear to be a better doctor, we do have a better judgment when somebody comes up with the sampling, sample data and sampling results. Thank you for today, and we'll see you in the next session.