And it's monitoring their reliability.

So first, reliability.

So, one, it will depend on the reliability of the, of the variables, so the reliability of x and the reliability of y.

We talked about validity, and reliability, and using NH, NHST as a way to get at reliability of correlations.

That's a pretty good reliability.

He has a good name for reliability.

I can assure you of his reliability.

That's it's, that's it's estimate of reliability.

And again nice and high again not great reliability estimates but, but good.

Matching the reliability, base load is actually irrelevant.

And in lecture five, we'll talk a lot about assessment of reliability.

If that, if that were a reliability, reliability estimate in my own research I would not wanna use that instrument again.

Cosby's argument concerns the reliability of the Hette Manuscript.

All he has going for him is his reliability.

So, in this first segment the focus will be on reliability and validity.

And if the correlation is high, that demonstrates some internal consistency or reliability.

The important topics to take away from this segment are validity and reliability.

I'm going to talk about three methods to estimate reliability.

The reliability of the device has not been the issue.

Just run the correlation between X1 and X2, and that provides you with a reliability estimate.

Just remember the magnitude of this correlation will depend on reliability. Easy way to remember that is the reliability of X refers to inaudible .

We'll walk through that. That's how we get estimates of reliability.

We are sceptical ourselves and we are running crosschecking routines to determine reliability of this conclusion.

In other words, reliability is impossible unless there is a natural warmth.

Microscopy, the standard WHO procedure, reaches from 40 to 60 percent reliability.

So, the magnitude of correlations critically depends on the reliability of our measures.

That's the inter item method and one of the most common ways we, we estimate reliability in psychology.

That's OK, Chelsea, you need to have six cars. Energy is all about reliability.

So, the fact the reliability estimate is this high, I was, I was impressed.

That's its maximum. So the high correlation you can get is the reliability estimate.

We need solutions either one or several that have unbelievable scale and unbelievable reliability, and, although there's many directions people are seeking,

I don't doubt the reliability of his information because it comes from a trustworthy person.

And then run a correlation analysis on those data and, again, if the correlation is high that demonstrates that we have some reliability.

And, what we're going to do is use that same example and do a test, retest reliability analysis as we talked about in

I introduced NHST in sort of bare bones fashion when we talked about correlation and the assessment of reliability in correlation.

And that's because the use of parallel tests and this idea of test retest, they are very simple to understand and, and provide really nice reliability estimates.

If, if a, if a measure X doesn't correlate with itself, meaning, it shows low reliability.

First we'll talk about reliability and validity in individual variables and individual, individual measures of psychological constructs, and then we'll talk about the idea of sampling.

So there's a way to get at reliability. We could also apply it to comparing two correlations.

On the one hand, our need for security, for predictability, for safety, for dependability, for reliability, for permanence.

And so this distance is called the margin of the support vector machine and this gives the SVM a certain robustness, because it tries to separate the data with as a large a margin as possible.

So we'll use this idea of classical test theory and these different methods of getting reliability estimates to make sure we have reliable instruments and reliable scores.

So, the goal for this segment is to follow along and write a script where we do a test, retest reliability and estimate analysis by running correlations.

But remember we are assuming univariate normal distributions, we're assuming linear relationships among variables, we are assuming homoscedasticity, reliability of validity and so on.

And it depends on this assumption of homoscedasticity Last thing I wanna do in this lecture is talk about the reliability of a correlation.