CSV Export from Workbench

From Cynefin.net
Revision as of 14:08, 12 February 2021 by Pstanbridge (talk | contribs) (Continual update of this page creation)
Jump to navigation Jump to search

The Workbench CSV export provides a number of different layout options.

This page deals with the data content and its interpretation, not the instructions on how to run an export from the Workbench tool. For that, see All Signifier module page.

Column Guide

FragmentID

This is the system identifier for the particular fragment capture and thus acts as the main key to each data record. It is a Universally unique identifier (UUID) Universally unique identifier.

NarrID

This is an ID that is generated with each export generation and can be used as a unique identifier whenever it is not important for the same ID to be associated with an entry in subsequent export generation. It exists as an historical item but some users find it helpful as they are easier to visualise than the UUID.

Session Start Date

This is the date when the respondent started to enter their fragment and signification process for a given fragment. It is often not present (why???). See date time zones

Session End Date

his is the date that the respondent pressed the save button for the fragment being entered. See date time zones

Server Entry Date

This is the date that the fragment was successfully updated to the server. See date time zones

Triad Columns

Each triad is output in either 5 or 6 columns. If all 5 or 6 columns are blank, then the triad was optional and the respondent did not select this triad but passed it by.

Each triad has what we call 3 anchors. The anchors are the labels at the triad corners or vertices. Triad data represents compositional data, thus the 3 anchor values together are not represented in Cartesian Coordinates but in what is called the Simplex Coordinate system. This is important because it means that normal multi-variate statistical methods cannot be used when analysing triad data. Thus we have the Left, Top and Right anchors for a triad.

See John Aitchison's | A Concise Guide to Compositional Data Analysis for details.

These are:

Triad Left Value

The value between 0 and 100 of the proportion of the triad value associated with the left anchor.

Triad Top Value

The value between 0 and 100 of the proportion of the triad value associated with the top anchor.

Triad Right Value

The value between 0 and 100 of the proportion of the triad value associated with the right anchor.

Triad X and Y Value

Two columns, one with the x value and one with the y value between 0 and 1. This is the actual 2D plot value. If you do a simple point plot of the X and Y values (provided you scale to an equilateral triangle) you will replicate the values as entered by the respondents.

Triad NA column

If the triad allowed for an NA checkbox, thus the respondent could declare that the triad was not applicable, then the 6th triad column is the NA column. It might not be labeled NA as SenseMaker enable the user designing an instrument to provide their own text label for this checkbox. The CSV output will give the integer "1" if the respondent had clicked the NA. If so, the other 5 triad columns will be blank.

Dyad Columns

Dyads are also compositional, but with two components and not three as with the triads.

Dyad Left Value

The value between 0 and 100 of the proportion of the dyad value associated with the left anchor.

Dyad Right Value

The value between 0 and 100 of the proportion of the dyad value associated with the right anchor.

Dyad X Value

The value between 0 and 1 of the actual x label value entered.

A little explanation: If the Left Anchor Value is 20 then the Right Anchor value is going to be 80. But the Dyad X value is going to be 0.8 as it represents the pointer placed toward the right of the slider bar (giving it a high right hand side label value than a left hand side label value).

Thus when plotting a histogram of the dyad values, simply plot the X value.

Dyad NA column

If the dyad allowed for an NA checkbox, thus the respondent could declare that the dyad was not applicable, then the 5th dyad column is the NA column. It might not be labeled NA as SenseMaker enable the user designing an instrument to provide their own text label for this checkbox. The CSV output will give the integer "1" if the respondent had clicked the NA. If so, the other 4 dyad columns will be blank.