From this webpage you can download the data from an online elicitation experiment on the use of quantified referring expressions designed and conducted by Dale Barr at the University of Glasgow, Kees van Deemter at the University of Aberdeen and Raquel Fernández at the University of Amsterdam.

Paper

The experiment and the data are described in the following paper:

D. Barr, K. van Deemter, and R. Fernández. Generation of Quantified Referring Expressions: Evidence from Experimental Data, The 14th European Workshop on Natural Language Generation (ENLG 2013), Sofia, Bulgaria, 2013.

Data

We make available the experiment instructions, anonymous metadata about the participants, the raw expressions produced by the subjects, and our classification of those expression into different classes of quantified referring expressions (see Table 1 in the paper).

• Instructions and debriefing: here you can see the instructions given to the participants before starting the experiment and the debriefing shown after they had finished.
• Data: all the collected data in XML format. This includes metadata on participant gender and native language (each participant is identified by a "SessionID" number), as well as the actual expressions ("Response") produced by each subject for each stimulus ("ItemID"). The data (including the stimuli) can more easily be visualised on this PDF file.
• Annotated corpus: this file contains a classification of the produced expressions into one of six categories: absolute (ABS), numeric (NUM), vague or base (BASE), comparative (COMP), superlative (SUP), and other (OTH). See the paper for further explanation. Note that this file does not show the expressions themselves but only the "SessionID" and the "ItemID", which can be matched to a "Response" in the XML file.

The data is freely available for research purposes. If you decide to use it, please send us an email to let us know and acknowledge our work by citing the paper above. Many thanks for your cooperation!

Acknowledgements

This research has been supported by the European Science Foundation EURO-XPRAG scheme.