## Do Justice and Trust Affect Policy Acceptability? Rizqy **Amelia** Zein <!-- Put the link to this slide here so people can follow --> article: https://doi.org/10.1108/IJHG-05-2019-0028 supp. materials: https://osf.io/gsd5t/ --- ## Aim <div style="text-align: left"> * [*Jaminan Kesehatan Nasional* (JKN)](https://www.thejakartapost.com/academia/2018/04/06/qa-bpjs-kesehatan-health-for-all-indonesians.html) is a national health insurance programme in Indonesia, covering around 221 millions Indonesians (2019). * The research aimed to examine variables that explain why ==laypeople== and ==healthcare workers== **support or disprove the policy**. </div> --- ## Hypotheses :one: <div style="text-align: left"> * We expected to find evidence that participants with ==higher just-world bias== would be ==more likely to support== the policy. * It was also assumed that if participants ==evaluated the quality== of Indonesian healthcare services ==more positively==, they would have been ==more likely to support== the policy. </div> --- ## Hypotheses :two: <div style="text-align: left"> * Participants who were ==confident== that ==health system and institutions== serve the public's interests, would ==support== the policy. * Participants who had ==higher trust in healthcare services==, would be also more likely to ==retain positive acceptability==. * At last, participants who ==voted for== *Partai Demokrasi Indonesia-Perjuangan* (PDIP), the political party that initiated the bill, the ==more they supported== the policy. </div> --- ## Methods :one: <div style="text-align: left"> * A cross-sectional survey involving 308 laypeople (:woman:=63.6%) and 95 healthcare workers (:woman:=64.2%). * We measured policy acceptability ($\omega$_t=.79), BJW ($\omega$_t=.84), trust in healthcare services ($\omega$_t=.90), and level of confidence to healthcare institutions ($\omega$_t=.82), support to a political party, and a general evaluation of healthcare service quality after JKN. </div> --- ## Methods :two: <div style="text-aligned: left"> * We tested the hypotheses using ==a two-level random-intercept linear mixed model== and built laypeople and HCW model separately * ==Laypeople model== :arrow_right: JKN member | non-JKN member * ==HCW model== :arrow_right: Doctor (GP, specialists, & dentist) | Non-Doctor (nurses, midwives, pharmacists, psychologists, etc) </div> --- ## Methods :three: <div style="text-aligned: left"> * All predictors were centred at the group mean * Estimating model parameters :arrow_right: maximum likelihood * Estimating degree of freedom :arrow_right: Satterwaithe approximation * For **HCW sample**, ==support for political party was omitted== from the model due to high percentage of **missing data** </div> --- ## Findings :one: <div style="text-aligned: left"> * **In laypeople sample**, we found ==a significant yet small== (*B*=0.043, *SE*=0.011, *p*=.000) effect of **trust** and ==large effect== of **general evaluation** of healthcare service after JKN (*B*=2.223, *SE*=0.255, *p*=.000) in explaining policy support. * We ==failed to reject *H*₀== that BJW, political party support, and confidence to healthcare system and institution were not correlated with policy support. </div> --- ## Findings :two: <div style="text-aligned: left"> * **In HCW sample**, we found ==a small and negative effect== of BJW (*B*=-0.164, *SE*=0.056, *p*=.004) and ==a positive, small effect== of **level of confidence** (*B*=0.536, *SE*=0.138, *p*=.000), and ==a moderate effect== of **overall evaluation** (*B*=1.521, *SE*=0.365, *p*=.000) to policy support. * We **failed to find an evidence** to conclude a relationship between **trust** and policy support. </div> --- ## Conclusion :one: <div style="text-aligned: left"> * Contrasting to our hypotheses, **HCW** with **higher just-world bias** tend to hold **more negative attitudes towards the policy**. * It's possible due to HCWs' perception that the system treats them unfairly :arrow_right: although further research is required. </div> --- ## Conclusion :two: <div style="text-aligned: left"> * Laypeople are **more likely** to base their judgement on the policy on **the quality of healthcare service** that they receive, not on a certain **ideology or values** :arrow_right: more pragmatic than ideological. </div> --- ## Limitations <div style="text-aligned: left"> * Participants were recruited through social media and IMs :arrow_right: prone to *selection bias* * Oversampling female participants </div> --- ### Social interventions to debilitate norms violations on social media The effect of censorship and social exclusion Rizqy **Amelia** Zein <!-- Put the link to this slide here so people can follow --> --- ### What we have learned so far :one: <div style="text-align: left"> * Moderately withholding unwanted contents [might reduce](https://academic.oup.com/esr/article/34/3/223/4944213) the occurence of counter-norms behaviour on social media, but censorship is **a double-edged sword**. * Despite its function to define descriptive norms, it could [evoke societal](http://doi.wiley.com/10.1111/pops.12391) and [behavioural problems](https://doi.org/10.1007/s12144-019-00316-8) :arrow_right: inducing ==psychological reactance==. * [Emotions (fear and anger)](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4675534/) plays an important role in resulting reactance. </div> --- ## ...and the questions arise :one: <div style="text-align: left"> * Moderate censorship is apparently quite effective, but what ==kind of censorship== specifically that works better in debilitating norms violations? * Could it be depending on who decide what to censor? :arrow_right: actors * ..or perhaps depending on the contents? * What are contents that "morally permissible" to be hidden? :arrow_right: the role of [ideology (RWA)](https://doi.org/10.1111/pops.12313) </div> --- ## Predictions :one: <div style= "text-align: left"> * Who decide what to censor * **Anger** might ==mediate== the ==effect of censorship== that comes from **a powerful source** (e.g. government) on ==increasing== **counter-norms behaviours**. * **Fear** (of social rejection) would explain why censorship leads to ==a decrease of norms violations==, esp when the contents are 'flagged' by **real users, or AI** (such as [@BotSentinel](https://botsentinel.com/)). </div> --- ## Predictions :one: <div style= "text-align: left"> * What contents are 'moral to be hidden'? * Contents from ==untrustworthy sources== and targetting ==vulnerable groups== would be perceived as more =="morally-permissible"== to be removed. * **Ideology** (*trait-like* disposition) * People with ==high RWA== would respond ==more negatively== if they suspect comments/contents that ==align to their values== have been deliberately ==removed==. </div> --- ### What we have learned so far :two: <div style="text-align: left"> * Don't feed the trolls :arrow_right: trolls [spend more time online](https://linkinghub.elsevier.com/retrieve/pii/S0191886914000324) and associated with [attention-conflict seeking](https://doi.org/10.1111/bjop.12154). * Therefore, [greater visibility](https://doi.org/10.1080/01972243.2017.1391911) is important to trolls. * Why don't we ==troll the trolls==? * If visibility is important, ostracizing is a potential punitive measure to impose norms. </div> --- ### What we have learned so far :two: <div style="text-align: left"> * Ostracism causes [extreme negative emotions](https://journals.sagepub.com/doi/full/10.1177/0022022117724900), but varying across culture :arrow_right: depending on self-construal and [personality](https://journals.sagepub.com/doi/full/10.1177/0146167216643933). * People are more [sensitive to social cues](https://journals.sagepub.com/doi/full/10.1177/1368430215596073) and [conform](https://doi.org/10.1007/978-3-319-16999-6_1474-1) to the group norms after being excluded. * The absence of paralinguistic digital affordances (PDA, e.g. <i class="fa fa-heart"></i>, <i class="fa fa-retweet"></i>, <i class="fa fa-thumbs-up"></i>, <i class="fa fa-share"></i>) [implies cyberostracism](https://doi.org/10.1177/2056305118800309) as users perceive it as a form of [social confirmation](https://www.ncbi.nlm.nih.gov/pubmed/27635443). </div> --- ## Predictions :two: <div style="text-align: left"> * **Ostracised individuals** might be ==more likely== to perform **morally-conformed expression** (<i class="fa fa-retweet"></i>, <i class="fa fa-share"></i>+ moral comments), especially after being rejected by **their ingroups**. * The effect would be ==mediated== by **anger** (independent self-construal) and **sadness** (interdependent self-construal), and ==moderated== by **personality** (agreeableness). </div> --- ## Proposed methods :one: <div style="text-align: left"> * Experimental approach is suitable to test proposed hypotheses :arrow_right: between-subjects design is ideal, if feasible. * To manipulate censorship, using 'fake' social media (e.g. [Truman](https://github.com/cornellsml/truman)) is ideal, but requiring heavy programming skill :arrow_right: potential technical and budget constraints. * Easier, yet less powerful alternatives :arrow_right: simulating online forum or using screenshots as stimulus. </div> --- ## Proposed methods :two: <div style="text-align: left"> * [Social media ostracism paradigm](http://smpo.github.io/socialmedia/) is a more powerful tool to manipulate ostracism, than the classic [Cyberball](https://www1.psych.purdue.edu/f~willia55/Announce/cyberball.htm) :arrow_right: fits very well to research aims and easily customisable. * [Registered report](https://www.nature.com/articles/d41586-019-02674-6) format is worth to consider :arrow_right: esp if aiming for high-tier outlets. </div> --- ### Thank you! :tada: You can find me on - [My personal website](https://rameliaz.github.io/) - [Twitter](https://twitter.com/ameliazein) - [or email](mailto:amelia.zein@psikologi.unair.ac.id)