Statistical exports

Answers become useful once they are out of the platform and inside the tool the analysis happens in. Forms offers five downloads, all from the responses page. Each needs permission to view forms, and each is written to your workspace activity log.

The spreadsheet table

The first export is the plain one: a table with a row for each response. Its columns run Response ID, Form Version, Individual ID, Full Name, Gender, Source, Submitted At and Linked, then one column for each question in the form, then any repeat group columns the form produced.

The same table is offered twice: as CSV, or as an Excel workbook whose header row is styled and frozen so that it stays in view as you scroll. The CSV carries a byte order mark, which is what makes Arabic answers open correctly in Excel rather than as strings of question marks.

SPSS, Stata and the metrics file

Two of the downloads are written for statistics packages, with one row per completed response in each.

The first is written for SPSS and carries the extension .sav

In it, a choice answer is stored as a number with value labels attached, so the package shows you the choice text rather than the code behind it. Yes and No become 1 and 0, labelled. A skipped answer is written as system missing rather than as a blank or a zero, which keeps it out of your averages. Each column label holds the full question text, cut at 255 characters, which is the length SPSS allows.

The second is written for Stata and carries the extension .dta

It has the same columns, and its labels are cut at 80 characters, which is the length Stata allows. One difference matters: the Stata file carries no value labels. Only the SPSS file has them, so with Stata you need the codebook beside you to know what a 3 in a column means.

A metrics file, also CSV, holds the computed figures rather than the raw answers: the scores and the gain figures the platform worked out for each response.

How the columns are named

One thing to know before you open any of them. In the SPSS file, the Stata file and the codebook, the columns are not named after the variable name you chose in the builder. They are named from the question's own number in the form, so they run q001, q002 and onwards, in the order the questions appear.

The codebook uses the same naming, which is what makes it usable: read the two side by side and every column has a meaning. Your own variable name is not lost either, because the codebook lists it against each column.

The codebook

The codebook is a CSV with one row per question, and it is the document you hand to a researcher along with the data. Each row gives the variable name, the question's order in the form, the question text, the Arabic question text, the question type, whether the question was required, its section, the valid values, the English value labels, the Arabic value labels and any notes.

It follows the labelling conventions the R haven package expects, so a codebook and a data file can be read into R together without anybody rewriting either of them by hand.

Opening the files

The SPSS and Stata files open directly in those packages, with no import step and no conversion to do. Open the codebook beside the data and a researcher who has never seen your form can tell what every column holds.

If the libraries the platform uses to write those two formats are not available, the download tells you that a dependency is missing rather than handing you a file that only looks right.