Solution to the 99 LISP Problems #3:

1
2
3
4
5
6

(defun element-at (alist n)
  (if (< (length alist) n)
    nil
    (if (= n 1)
      (car alist)
      (element-at (cdr alist) (1- n)))))

Lisp dialect: Steel Bank Common Lisp

Solution to the 99 LISP Problems #2:

1
2
3
4
5

(defun my-but-last (alist)
  (if
    (<= (length alist) 2)
    alist
    (my-but-last (cdr alist))))

Lisp dialect: Steel Bank Common Lisp

Solution to the 99 LISP Problems #1:

1
2
3
4
5
6
7
8

(defun my-last (alist)
  (if
    (equalp (length alist) 0)
    nil
    (if
      (equalp (length alist) 1)
      alist
      (my-last (cdr alist)))))

Lisp dialect: Steel Bank Common Lisp

In the last post, I talked about the test of statistical significance. Here I talk about the next major step in frequentist inference, i.e. Neyman Pearson hypothesis testing. I will also attempt to simultaneously elicit the differences between the two.

Prologue

The past couple of days I have been reading about tests of statistical significance and the associated maladies since I came across this paper shared on a social networking site.

In grad school conferences, I used to hear myriad tales of caution about p-values being unreliable, prone to misinterpretation and so on. Yet, no econometrics prof will touch this topic in class with a barge-pole. By the second semester, with term papers becoming the norm, the whole class was hunting for stars in the NHFS wilderness. PhDs were always discussing how they spent the whole night trying to find a model that will render the variable of interest (on which their thesis rested) significant at 5% level. I did too. We were taught to.