6
$\begingroup$

Why do a pair of variables with no significant correlation and no significant regression intercept and slope, have a highly significant regression with high adjusted $R^2$ when the regression is forced to pass through the origin?

$\endgroup$
6
  • 1
    $\begingroup$ Possible duplicate of What are the uses and pitfalls of regression through the origin? $\endgroup$ Commented Dec 8, 2016 at 12:35
  • $\begingroup$ This doesn't quite seem like a duplicate to me. $\endgroup$ Commented Dec 8, 2016 at 13:22
  • $\begingroup$ @gung, is it customary to bump ancient threads with minor edits like the few you did recently? I noticed kjetil b halvorsen doing the same thing. I'm asking sincerely and not trying to be rude. $\endgroup$
    – gammer
    Commented Feb 2, 2017 at 3:26
  • $\begingroup$ @gammer, I don't know how to evaluate if it's "customary". We do occasionally edit / bump old threads--there isn't a rule against it. I believe that having threads appropriately tagged is very important to the organization of the material on the site. But when I do that, I only do a max of 5-6 at a time, & generally when traffic is slow. $\endgroup$ Commented Feb 2, 2017 at 3:37
  • 1
    $\begingroup$ I wouldn't bump an old thread just to remove "thanks", @gammer. But if I do edit a thread, I try to get all that stuff at the same time. $\endgroup$ Commented Feb 2, 2017 at 3:46

2 Answers 2

15
$\begingroup$

Here is an illustration that simulates $y$ and $x$ independently of each other so that the true slope is zero. The mean of $y$ is nonzero, such that the true intercept is also nonzero.

The LS line without intercept must start at $(0,0)$ without intercept, and will try to "catch up" with the data points as quickly as possible if $y$ has nonzero mean, which induces a clear slope (purple line), while the blue line with intercept may start at the right level for $y$ right away, such that it "needs" no slope.

Note however that this example will typically exhibit a significant intercept in the model with intercept.

enter image description here

n  <- 100
mu <- 10
y  <- rnorm(n, mean=mu)
x  <- runif(n)

plot(x, y, ylim=c(0, mu+3))
abline(v=0, lty=2)
abline(h=0, lty=2)
abline(lm(y~x), col="lightblue", lwd=2)
abline(lm(y~x-1), col="purple", lwd=2)
abline(h=mu, lwd=2)
legend("bottom", legend=c("with intercept","without intercept","truth"), 
       col=c("lightblue","purple","black"), lty=1, lwd=2)

We can also analyze the issue theoretically. Suppose the true model is $$ y_i=\alpha+\epsilon_i, $$ i.e., $$ y_i=\alpha+\beta x_i+\epsilon_i\qquad\text{with}\qquad\beta=0 $$ or $E(y_i|x_i)=E(y_i)=\alpha$.

Under this model and assuming $E(x_i\epsilon_i)=0$ for simplicity (i.e. no further misspecification than a missing intercept), the plim for the OLS estimator $\hat\beta=\sum_ix_iy_i/\sum_ix_i^2$ of a regression of $y_i$ on $x_i$ without constant is given by \begin{align*} \text{plim}\frac{\sum_ix_iy_i}{\sum_ix_i^2}&=\text{plim}\frac{\sum_ix_i(\alpha+\epsilon_i)}{\sum_ix_i^2}\\ &=\text{plim}\frac{\frac{1}{n}\sum_ix_i(\alpha+\epsilon_i)}{\frac{1}{n}\sum_ix_i^2}\\ &=\text{plim}\frac{\alpha\frac{1}{n}\sum_ix_i+\frac{1}{n}\sum_ix_i\epsilon_i}{\frac{1}{n}\sum_ix_i^2}\\ &=\frac{\alpha E(x_i)}{E(x_i^2)} \end{align*} For example, in the numerical illustration, we have $\alpha=10$, $E(x_i)=1/2$ and $E(x_i^2)=1/3$.

Hence, unless we are in the special cases that $E(y_i)=0$ or $E(x_i)=0$, OLS is inconsistent for $\beta=0$, $\text{plim}\hat\beta\neq0$.

In the first case, we do not need a sloping $\hat\beta$ anyway, in the second, a flat line is "best" for OLS as smaller squared mistakes for positive fitted values for positive $x_i$ (in the case of a positive estimated slope) would be overcompensated by much larger squared mistakes for negative fitted values for negative $x_i$.

$\endgroup$
3
  • 3
    $\begingroup$ Nice illustration. $\endgroup$ Commented Dec 8, 2016 at 12:39
  • 1
    $\begingroup$ Are you sure your data have "no significant regression intercept"? $\endgroup$ Commented Dec 8, 2016 at 13:22
  • $\begingroup$ @gung, you are right, I edited my answer to highlight that the properties of my example. $\endgroup$ Commented Dec 8, 2016 at 13:51
-1
$\begingroup$

Basically, to force a regression through zero the statistical software will enter in an infinite amount of data points at (0,0). This makes the normal R^2 formula useless, and a different R^2 formula is used. The result of this different R^2 formula is always very high. You can go to this link to get more specifics- https://www.riinu.me/2014/08/why-does-linear-model-without-an-intercept-forced-through-the-origin-have-a-higher-r-squared-value-calculated-by-r/

$\endgroup$
1
  • $\begingroup$ I'm sorry, but I can't understand what you mean when you're referring to an infinite number of data points. Can you please explain it? $\endgroup$ Commented May 24, 2019 at 23:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.