Introduction to Sampling | Applied Biostatistics | BIO733_Topic059

Duration: 5:48 724 transcript words EN summary

Video summary

The module introduces the fundamental rationale behind obtaining a sample and outlines the procedures involved in its collection, highlighting both the advantages and disadvantages of this approach. Before delving into these details, it is essential to establish basic terminology, distinguishing between the population, which represents the totality of characteristics under study, and the census, which refers to studying the entire population. Within this framework, individual members are known as population elements, while the true values observed from them are called population parameters. Sampling serves as the process of selecting a subset of observations, referred to as a sample, to provide an adequate description and draw inferences about the larger population, making the sample the most representative part of that whole. A critical component ensuring the accuracy and reliability of sampling procedures is the sampling frame, defined as a complete list of all individuals included in the population from which the sample will be drawn. While the population encompasses everything we wish to discuss, the sampling frame specifically includes those elements available for study. The primary motivation for using samples rather than conducting a census on the entire population stems from practical limitations; populations are often too large, making a full census expensive, time-consuming, and prone to other logistical issues. Consequently, researchers aim to obtain a representative sample that allows them to speak economically about the population without needing to observe every single member, a process known as statistical inference when moving from sample observations back to population conclusions. Despite these benefits, sampling is not without its drawbacks, primarily revolving around variability and potential bias. Since samples can vary significantly from one another, two different samples might yield totally different answers or only slight variations, introducing uncertainty into the results. Furthermore, there is always a risk of introducing bias during data collection, which can stem from human error or the difficulty in securing a truly representative sample. In large-scale surveys, reliance on manpower becomes necessary, and if this workforce is not adequately trained, it can lead to further errors. Additionally, issues such as the absence of informants and various other types of errors can compromise the accuracy of the findings, making it challenging to ensure that the sample perfectly reflects the population it intends to represent.

Read the full video transcript

Hello. In this module, we will be talking about the rationale of obtaining a sample. We will further look at the procedure of collecting a sample as well as its advantages and disadvantages. Before we move further, we do want to talk about few basic terminology that includes the population, which is the totality of aggregate of characteristic that is under our study. Whereas, whenever we we want to study the population, the process or the study of the population is called census. In a population, there are individual observations. Each individual member that for which the characteristics we want to study is called the population element. Whereas, any value or characteristic that we observe from the population is called population parameter, which are the true values and they are constants. Hence, the sampling is a process of selecting the observations, which is called a sample, to provide an adequate description and inferences of the population. So, sample is the most representative part of the population. Whereas, the sampling units are the individual person, animal, or object that has the measurement or observation taken on them. But, very important feature that makes our sampling procedures much more accurate and reliable is the sampling frame. Whereas, the sampling frame is a complete list of all the individuals that are included in our population or from which we are going to collect our sample. So, as we know that we always want to talk about the population. Whereas the sampling frame is a part of our population which specifically we want to study or which are available for our sample. The idea is that we cannot study the population in all the situations. Because population is too large and to conduct the censors out of this population is always having a lot of other [snorts] drawbacks. Like it could be expensive. It could be time-consuming. And there are many other issues that can come in. So, to be more economical but still being able to talk about the population, we always try to take a representative part of the population which we call sample. The sampling is a process of obtaining that sample from the population. So, population is something that about which we want to talk about and sample is something that we actually observe to talk about the population. And when we talk, we go from the sample to the population, the whole process is called the statistical inference. Since we know we cannot study overall population in all these instances, therefore we rely always on a sample. And it is very important that our sample should be good. There are few advantages of the sampling that it is economical in nature. Because we don't have to study the overall population. Moreover if it's carefully done, then it gives us very accurate results regarding the population. Moreover, it also give us very reliable estimates for the population. It also saves time and only resort to the state if we are about to study an infinite population. Because practically it's really hard to study the whole population in in case of infinite populations. But as every method has its disadvantages, too. It Sometimes it could be inadequate. Because samples do vary from sample to sample. So, it's very likely the two samples may give us totally different answers. Or it's most likely that they are slightly different or even the same. There's always a chance of bringing some bias into the results when we are collecting the sample and that could be human error as well. Then there's also problem of accuracy in our results. And if we are taking the sample, there's Sometimes we get into the problem that it's very difficult to get the representative sample. Since for to to obtain a sample, we always rely on some manpower, so it's very likely that manpower is not very trained. Though for the smaller studies, a PI can himself take the sample. But if we are conducting our large-scale surveys, then we do need some manpower and it's very likely that manpower is not trained well. Then there it's very likely that our informants are absence. Moreover, there are chances of committing the errors in the sampling. Or there are different other type of errors that that can come in the play. Thank you.