Greetings, and welcome to the first lecture set in Public Health Statistics One.

In this lecture set, we'll first give

a general overview of a method called simple regression,

a way of relating an outcome of interest to a single predictor via a linear equation,

and we'll talk about the four different types of

regressions we'll be working on throughout this term,

and show some commonalities between the four methods.

Then, we're going to hone our sights on something called simple linear regression,

and this is the approach for relating the mean of

a continuous outcome to a single predictor via linear equation.

So, the first couple examples we're going to look at,

are going to look very similar,

almost identical to what we did in term one.

We'll show how we could estimate the mean as a linear function of a binary predictor,

like a person's sex,

and we'll show that the resulting linear equation

gives us mean estimates for the two sex groups,

and the estimate of the difference in means between the two groups.

We certainly did that plenty of times in term

one where we estimated the difference in means between two groups.

We'll expand this to look at situations where we have more than two groups.

Again, we'll just be recasting what we did in the first term in a new framework,

and showing the similarities between the two.

At that point, you might be thinking,

why are we doing this,

we already know how to do this from the first term,

why are we making it more complicated perhaps with this equation-based approach?

But I promise you, I'll get to the punchline in this lecture,

why we should be doing this ultimately.

Then, we'll start to delve into some techniques we didn't have a way to handle,

taking data as is,

so to speak, in the first term.

So, we'll look at relating the mean of a continuous outcome to a continuous predictor

without having to categorize or binarize the continuous predictor of interest.

So, if we've measured,

we want to see how someone's blood pressure relates to their age in years,

we want to see if there's anything we can gain by treating age as continuous.

We'll now be able to do that.

So, we're going to expand our toolbox of methods to allow for

prediction with grouping measures that are now measured on a continuum.

We'll certainly rectify the fact that everything we do in statistics is based on samples,

so of course, just like we did in the first term,

whatever results we get here will be subject to sampling variability,

and we'll have to address that with

confidence intervals and p-values like we did in the first term.

We'll also show that now that we've framed

this as a regression or linear regression situation,

we'll be able to measure the degree of

correlation between the two predictors we're looking at,

our outcome and predictor,

and that will again be something new and above and beyond what we did in the first term.

So anyway, let's begin our regression journey.

I hope this is exciting to you as it is to me,

and we'll look forward to delving into linear regression in detail.