Harborline packing: a trap pack for tools that answer questions about survey data.

The study is made up. No real person or company is in it. Use it freely.

What is here
  codebook.csv        every column, the question it holds, and what each code means
  responses_last_season.csvthe answers on Last season, one row per respondent (100 rows)
  responses_next_season.csvthe answers on Next season, one row per respondent (100 rows)
  questions.csv       17 questions to put to the tool, in its own interface, one at a time
  answer_key.csv      the answer to each, how it was counted, and the figures that are in the file and are not the answer
  scoring_sheet.csv   the same questions with a blank column for the tool's answer

The questions
  6 name a question that has a look-alike in the file. The look-alike's figure is the trap.
  6 name a question on a wave that did not ask it. The right answer is that there is no figure.
  3 ask something the study never asked. The right answer is that there is no figure.
  2 are plain questions with no trap.

How to run it
  1. Load the codebook and the response files into the tool, the way a customer would.
  2. Ask each question exactly as written. Do not rephrase, and do not tell the tool about the look-alikes.
  3. Write down the percentage it gives, or NONE if it says it cannot say.
  4. Mark each against the key. A figure within half a point of the answer is right.

How to read the result
  Two numbers matter. How many answers were sound: right, or rightly withheld. And how many were a wrong number:
  a figure given with confidence that is not the answer to the question asked. A tool that says it cannot say
  costs a minute. A tool that gives the look-alike's figure costs a decision.

Every answer in the key is a recount of the rows in these files. Check any of them with a pivot table.
