Let's say I have a sample size of 1,000 individuals. These individuals can easily be sorted into various groups. So, I ask them a question, say, "Do you understand how to label your values?"
This is how many people answered yes:
Those who passed Algebra with an A - 34.3
Those who failed - 69.7
Those who plan to retake Algebra - 28.9
Those who will never touch a math text again - 78.3
...I decided to do a little simple addition:
34.3 + 69.7 = 104.0
28.9 + 78.3 = 107.2
So, I ask you--since Ken Ham and his number crunching buddy and his editor couldn't be bothered to check his numbers or give me a label of what kind of number I'm looking at--what does it mean that 107.2 of my 1,000 said yes?
~Luke Holzmann
Filmmaker, Writer, Expectant Father
Subscribe to:
Post Comments (Atom)
5 comments:
I didn't follow your link this time, but did he forget to consider the statistical error, or do it incorrectly?
Maybe say those first set of numbers should have been 32.3 and 67.7 with a ± of 2?
Whatever the problem, a good editor/proof reader should have caught the math error.
I have to say that I think this one is just hosed. If the statistical errors were involved, usually that might give you an answer of 100.1 or 99.9. I think you're right: someone needs to check their math more carefully.
So let me ask you something: when you see errors like this in writing, do you find it harder to believe the point that the author is trying to make? I'm just asking because using messy data tends to make me rather skeptical. (And if it's a scientific journal I'm reading, I tend to blow off the author altogether.)
Luke, my thoughts exactly. It's very unclear what the numbers represent from the text and the diagram. Really, there's no excuse for it being like that.
However, I think, based on the fact that all the previous numbers in that section are percentages of the two groups (those who still attend on holidays, and those who never attend), it seems that the best interpretation of those numbers would be as similar percentages. Like all the previous (but labeled) tables in that section of the book (at least, what I can see on Google Books), I think the table should be interpreted this way:
34.3% of people who attend church on holidays said that they believe premarital sex is okay.
69.7% of people who never attend church said that they believe premarital sex is okay.
28.9% of people who plan on coming back to church said that they believe premarital sex is okay.
78.3% of people who are never coming back to church said that they believe premarital sex is okay.
Although they failed to sufficiently label their data on that table, I think the intended meaning can be fairly clearly deduced from the larger context.
I'd like to see the data. Depending on how he went about calculating the percentages, Ham might have messed up inputing the numbers. So for example, he might be looking at 130 out of 100 people because 100 said they were "against" premarital sex while 30 said they were "strongly against" it, and he forgot to add them. I've seen this sort of thing happen a million times. All someone need to remember to do is make sure everything adds up to 100%, but they don't think of that until they're in the middle of a presentation.
Second, cases could be misassigned
You are correct, it would have been better if the author, the editor, the illustrator, the typesetter, and/or the proofreader had caught the fact that the numbers really needed percent signs after them, but I think the meaning is quite clear from the context . . . and Abe has interpreted things correctly.
Post a Comment