Tipsters:
Question 1. Most texts do frequency distributions from highest to lowest
score, as in:
X tally
34 xx
33 xxxxx
32 xxxxxxx
.
Has there ever been any reason to list from high to low?? It seems
to me that you'd want to list from low to high, this is the custom for
other techniques such as stem-n-leafs and histograms. Indeed, if you are
going a tally, you can simply rotate the frequency histogram 90 deg
counterclockwise and get a histogram for no extra cost. A few books
listed from low to high (my preference) en route to a histogram - but
didn't call it a "frequency distribution."
Question 2. Some texts claim that when constructing grouped
frequency distributions, to make the stated lower limit of the first
interval evenly divisible by the interval width. So, if we had the
numbers 10,11, 12, 13, 14,14,14,15,16,16, and so forth, and chose an
interval width of 3, the intervals would be:
9-11
12-14
15-17 etcetera.
What is the reason for the rule? I can only come up with one. If your
numbers get doen to zero or less than zero, you will have an interval
that starts with zero. In the above example, if you continued downward
you'd get to 0 - 2 (However, if plotted the data you'd use midpoints,
you generally label the interval by its midpoints). I played with
Minitab a little I generated a few histograms that, by default, violated
the rule. One would think grouped histrograms would obey the same rule.
Is this simply an relic from the past - that when people did these by
hand they wanted a standard rule to guarantee some uniformity??
--
------------------------------------------------------------------
John W. Kulig [EMAIL PROTECTED]
Department of Psychology http://oz.plymouth.edu/~kulig
Plymouth State College tel: (603) 535-2468
Plymouth NH USA 03264 fax: (603) 535-2412
------------------------------------------------------------------
"One word of truth outweighs the whole world."
Russian proverb
------------------------------------------------------------------
---
You are currently subscribed to tips as: [EMAIL PROTECTED]
To unsubscribe send a blank email to [EMAIL PROTECTED]
