How to Lie with Statistics

I finally am satisfied that my new software works and is spitting out correct data. The hard part is interpreting the data. That is why they wrote that book on "How to Lie with Statistics", by Darrell Huff (Author), Irving Geis (Illustrator), © 1954 (Good Grief, the same year I was born!) Reprinted in paperback 1993, still used in many beginning statistics courses! My wife's intro course used it in 1972!

For example, 2,570 (25%) active Churches did not pay percap in 2007 and they had a combined income of $270,431,754 (13%)!, Makes for a good headline, but is it accurate? Which of the following numbers is the most meaningful to you?

Description	# Churches	Income	Percap	Members
Active Churches	10,172	$2,097,432,727	$43,424,221	2,104,855
No Percap 2007 raw	2,570	$270,431,754		381,134
No Percap 2007 NCD (non-reporting)	- 117	$0		0
No Percap 2004-2007 & (w&m&I>0) Non reporting?	- 530	$146,815,885		131,965
No Percap 2007 corrected	1,923	$123,615,869	$5,137,864 (Avg 20.62)	249,169
No Percap 2006 but some in 2007 & (w&m&I>0)	619	$140,542,085	$3,338,842	120,494
No Percap 2006-2007 but some in 2004/2005 & (w&m&I>0)	198	$51,870,420	$589,240 (2005)	35,017
No Percap 2007 but some in 2004-2006 & (w&m&I>0)	491	$109,412,120	$1,869,013 (2006)	90,042

In trying to figure out how many churches failed to report for 2007, 765 Churches have m6=m7>0 and w=0 and i=0 and p=0 AND 64 Churches have m=0 and w=0 for the first time.

See my other page Statistics for 2007 for more lies, .... data!