A dataset of 108 novel noun-noun compound words with active and passive interpretation

We created a dataset of 205 English novel noun-noun compounds (NNCs, e.g., “doctor charity”) by combining nouns with higher and lower agentivity (i.e., the probability of being an agent in a sentence). We collected active and passive interpretations of NNCs from a group of 58 English native speakers. We then measured interpretation time differences between NNCs with active and passive interpretations (i.e., 108 NNCs), using data obtained from a group of 68 English native speakers. Data were collected online using crowdsourcing platforms (SONA and Prolific). The datasets are available at osf.io/gvc2w/ and can be used to address questions about semantic and syntactic composition.