Question

In your own words, explain what sampling error is.

Why is sampling error such an issue when it comes to inferential statistics.

What is alpha? What does it represent in hypothesis testing?

Now that you know a little more about hypothesis testing, how do you feel about the fact that hypothesis testing will never give you a certain answer—that there’s always a possibility of creating a Type I or Type II error?

Answer 1

Answer to the question)

Sampling error: are the errors that originate while taking the sample, or recording the sample responses

Now when one needs to draw and inference or conclusion related to the population, it is important that the sample represents the true population correctly

if this is not the case , if the sample does not represent the true population, any conclusion you draw based on the sample would not be true for the population

Hence the sole purpose of drawing inference about the true population is not served just because of sampling error

Thus sampling errors are important as they hinder in the functioning of inferential statistics

.

Alpha is the probability of type I error that the researcher is willing to take during the research process. This probability of type I error is set before the research starts. For example when we say alpha = 0.05 , this implies that the researcher is willing to take 5% chance of type I error

Now since alpha is set, it finally helps in deciding if the null has to be rejected or not

The P value obtained is compared to this alpha, and that is the reason why it is also called level of significance

.

Since we are dealing with statistics, and relates to future inference. Future is always uncertain. We can predict, draw inferences and forecasts, but that just provides us a predictable path, it doesnot guarantee what is going to happen in future

Likewise when hypothesis testing is conducted, the conclusion draw is valid in general, except a few scenarios where errors might occur

But the usefulness of the process of hypothesis testing outnumbers the chances of error, hence proving it to be an important tool in inferential statistics