We don’t know which is correct - or if another model is better. We can only examine correlation.
Describes the relationship between two variables. Not scaled.
\(\sigma_{xy}\) = population level covariance
\(s_{xy}\) = covariance in your sample
\[\sigma_{XY} = \frac{\sum (X-\bar{X})(y-\bar{Y})}{n-1}\]
\[\Large\rho_{xy} = \frac{\sigma_{xy}}{\sigma_{x}\sigma_{y}}\]
Y is perfectly predicted by X if r = -1 or 1.
\(r^2\) = the porportion of variation in y explained by x
Ho is r=0. Ha is r \(\ne\) 0.
Testing: \(t= \frac{r}{SE_{r}}\) with df=n-2
WHY n-2?
\(\sigma_{xy}\) Because you use two parameters: \(\bar{X}\) and \(\bar{Y}\)
\[SE_{r} = \sqrt{\frac{1-r^2}{n-2}}\]
inbreeding.coefficient pups
inbreeding.coefficient 0.01 -0.11
pups -0.11 3.52
inbreeding.coefficient pups
inbreeding.coefficient 1.00 -0.61
pups -0.61 1.00
estimate | statistic | p.value | parameter | |
---|---|---|---|---|
-0.608 | -3.589 | 0.002 | 22 | |
</div | > |
Transform variables to ranks, i.e.,2,3… (rank()
)
Compute correlation using ranks as data
If n \(\le\) 100, use Spearman Rank Correlation table
If n \(>\) 100, use t-test as in Pearson correlation
Then it’s code in the data, give the keyboard a punch
Then cross-correlate and break for some lunch
Correlate, tabulate, process and screen
Program, printout, regress to the mean
-White Coller Holler by Nigel Russell
\(\widehat{Y} = \beta_0 + \beta_1 X + \epsilon\) where \(\beta_0\) = intercept, \(\beta_1\) = slope
Minimize Residuals defined as \(SS_{residuals} = \sum(Y_{i} - \widehat{Y})^2\)
\(\LARGE b=\frac{s_{xy}}{s_{x}^2}\) \(= \frac{cov(x,y)}{var(x)}\)
\(\LARGE = r_{xy}\frac{s_{y}}{s_{x}}\)
Least squares regression line always goes through the mean of X and Y
\(\Large \bar{Y} = \beta_0 + \beta_1 \bar{X}\)
\(\Large \beta_0 = \bar{Y} - \beta_1 \bar{X}\)
Data Generating Process:
\[Visits \sim Resemblance\]Error Generating Process:
\[\Large Visits_i = \beta_0 + \beta_1 Resemblance_i + \epsilon_i\]
\[\Large \epsilon_i ~ N(0, \sigma)\]
Points fall on 1:1 line, no systematic deviations
No systematic trends in relationship required!
Appears peaked in the middle…
Nothing with > 1
Should be a cloud with no trend