How to interpolate for degrees of freedom from a $ \chi^{2} $ table [duplicate]

Question

The question asks to find $ \chi^{2*} $ that cuts off 5% of the tail with df=52. the chart given lists df=50 then jumps to 60

You may want to remove the * from your body text as well. Your question is covered in the answer here: How do I find values not given in (interpolate in) statistical tables? — Glen_b, Commented Nov 25, 2015 at 21:53
in the title it was an error, in the question it just means a specific value of $\chi^2$ opposed to the whole range — Mitch Hughes, Commented Nov 25, 2015 at 21:56
It's probably more confusing than enlightening. I'd suggest "find the $\chi^2$ value that..." — Glen_b, Commented Nov 25, 2015 at 21:58
Specifically, note that for degrees of freedom for $t$, $\chi^2$ and $F$ the linked answer recommends linear interpolation in $1/\text{df}$ rather than in $\text{df}$. However in this particular case, linear interpolation in $\text{df}$ is actually very good, slightly better than inverse interpolation in df; I'd stick with that for this case (which is what the first half of the linked post is about). — Glen_b, Commented Nov 25, 2015 at 22:09

Espe FG · Accepted Answer · 2015-11-25 21:58:06Z

-2

I would take the data from df=50 in this case. If it was 55--> average

answered Nov 25, 2015 at 21:58

Espe FG

11 bronze badge

$\begingroup$ (-1) This is not interpolation at all. If you're going to average for $df=55$, which is (linear) interpolation, then why not linearly interpolate for $df=52$? Why introduce such an inconsistency, especially when the error made when replacing $52$ by $50$ is substantial? $\endgroup$
– whuber ♦
Commented Nov 25, 2015 at 22:03

Add a comment |

1 Answer 1