Replication Data and Analysis Code for: Goal-Oriented Modeling and Analysis of Explanation Requirements

Replication data of our experimental study evaluating a framework for capturing and implementing explanation requirements.

The data consist of 20 data points from an equal number of participants. The participants were presented two cases of software-intensive systems: a socio-technical system, namely a University Petitions case, and an AI-based system, namely Home Automation case. They performed two main tasks. Firstly, they were presented with logs from activity within the context of those systems. They were asked which of the listed actions are explanation-worthy / answering on an 5-point ordinal scale. Secondly, for selected actions, they were then presented with 2-4 statements meant to explain the action. They where asked to rank the explanations and also rate them (5-point ordinal scale) with respect to the usefulness of the explanation.

This data package contains the responses (data.xlsx and data.cvs) to be interpreted based on survey.txt, the psytoolkit script used to generate the survey. An Analysis.Rmd script performs the data analysis presented in the paper.