Abstract
New Zealand’s 2020 cannabis legalisation referendum was narrowly defeated. Some post-referendum commentary claimed that the anti-reform campaign dominated the information space. Digital media represented a new forum for referendum advocacy. We analysed reporting on the cannabis referendum by six leading New Zealand digital news providers and Facebook advocacy campaigns ‘for’ and ‘against’ legalisation during 3 months pre-referendum. The mean sentiment score of media articles was marginally supportive of reform (+0.4 on a scale of −2 to +2). On average, pro-legalisation articles were re-published more often (2.3 vs 1.5 times for anti-reform), received better website placement (52% of articles published in the prime website section) and had more Facebook interactions (mean 1129 vs 771 for anti-reform). Ninety six per cent of articles were shared on Facebook, including via paid advertising. The principal pro-legalisation campaigner spent four times as much on Facebook advertising as the principal anti-legalisation campaigner. The networked map of Facebook posts illuminated links between digital media and cannabis advocacy. We find that pro-legalisation campaign dominated the digital information space, while the anti-legalisation campaign relied on traditional media, such as billboards. Combining media sentiment analysis with network analysis of social media posts provided new insights into the cannabis referenda debate.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 Data collection in the afternoon and evening was employed in order to account for the typical increase in Internet traffic in the latter half of the day (e.g. Morley et al., 2018).
2 The category of “political and social issues” advertisements was first introduced by Facebook in 2018 to provide greater transparency on advertising related to elections (Facebook, 2019). The feature was rolled out in New Zealand mid-July 2020.
3 Crowdtangle™ is Facebook-owned social media monitoring software.
4 Facebook Ad Library only provides a range estimate for spending (example: $500–600) and impressions (example: 45–50,000) for individual advertisements in the library, not the exact values. Median values for individual ads were used in calculations.