The Statistics Lesson

"I just don't get statistics," Katie said plaintively.

She and I and Luke were taking a graduate level seminar on statistics in archaeology, and the fact that Katie didn't get statistics was not news to me and Luke. We had already gone over and over Pearson's r and the student's t, correlation coefficient, nearest neighbor and random sampling, and still she wasn't catching on at all.

You have to understand--statistics were very important to archaeology in those days. During the late 70s and early 1980s, archaeology went through a technical spasm where the only acceptable paper published nearly anywhere had to have Greek letters scattered throughout it. This was never more the case than in the graduate seminar we took, taught by one of archaeo-statistics' main practitioners.

Luke and I grimaced at each other. Why did we care whether this dope got statistics or not? Unfortunately, we had been given a joint assignment, and each of our grades would be contingent on the performance of the others. Truth to tell, under any other circumstance, we'd have preferred our fellow students to get much worse grades than us; but in this course now we had to work to bring Katie up to speed.

Luke said, "Look, Katie, it's not that hard. All statistics is--all those little Greek symbols are--is a mathematical restatement of a sentence."

"Huh?"

"So, you see," said I, bravely taking up the cudgel. "Say you've excavated this site, for instance, and it had 200 cache pits on it."

"Oh, yeah, that's good; and say you want to compare the contents of the cache pits, to see if there's any pattern to them."

"Pattern?"

"Yep," said I. "Say you think maybe you noticed that while you were excavating the east side of the site, you noticed more, uh, marine shell coming out of those pits."

"And you thought it might be interesting to see if there were other patterns of distribution," chimed in Luke. "So you'd run a statistic to see if there was a difference between the pits, you know, if the marine shell really did come out of the eastern pits."

Katie wrinkled her forehead. "Why not just make a series of maps, code the cache pits by their contents?"

"Well, yeah, that'd work, but with statistics you get a number for comparison."

"A number?"

"Yeah, you know, a probability figure, a number which represents the probability that marine shell fragments come from eastern most pits."

There was a brief pause. "What good is a number?"

"What?"

"But what good is a number? It doesn't really show you what you want to know--which is where are the marine shell."

"Well, sure it does, kinda. A number allows you to compare one set of data to another. So, say you could run the same statistic for a different kind of artifact, and you could then say that the marine shell showed up more frequently in the eastern pits than, say, the freshwater shell did. And you could put a number to the difference between the two."

Katie raised an eyebrow. "And of course, this is a meaningful number?"

"Sure, it's meaningful," said Luke, "as long as you carefully collect and input the data."

"Because, of course," Katie said triumphantly, "we always know that we've got all the pertinent data."

Luke's teeth began to grind together.

"I still think a map would show that more clearly and reasonably."

Another pause. "Luke, I think she's starting to convince me."

---

Although this conversation (or a reasonable facsimile thereof) took place in the 1980s, trying to develop a reasonably descriptive statistical technique continues in the profession. Page 2, 3, and 4 of this article list recent and not-so-recent articles on the use of statistics in archaeology.

Previous

• 1
• 2
• 3
• 4

Next