The Effects of WhatsApp on Politics: A Multi-Country Deactivation Experiment

class: center, middle, inverse, title-slide

.title[
# <span style="font-size: 80%;">The Effects of WhatsApp on Politics: A Multi-Country Deactivation Experiment </span>
]
.author[
### Tiago Ventura | Georgetown University
]
.date[
### <span style="font-size: 80%;"> <br> Tech & Public Policy Grantee Conference <br> 11/14/2023 </span>
]

---

layout: true

<div class="my-footer"><span>Tiago Ventura | Georgetown University &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp &nbsp Tech & Public Policy Grantee Conference </span></div>

---

class:middle
### Motivation

.center[
<img src="nyt.png" width="100%" />

]

---
class:middle

### WhatsApp Multimedia Deactivation Experiment - Brazil 2022

- Incentivized participants to spent three weeks without accessing any multimedia (images, videos and audio) on WhatsApp.

---
class:middle

## Intervention

.pull-left[
.center[
<img src="./output/iphone_deact.jpeg" width="70%" />
]
]

.pull-right[
<br>
.center[
<img src="./blur.jpg" width="80%" />
]
]

---
class:middle, inverse, center

# Results

---
class:middle

### Effects on Exposure to Misinformation Rumors (H1)

.center[
<img src="./output/exposure_h1.png" width="90%" />
]

---
class:middle

### Effects on Belief Accuracy (H2)

.center[
<img src="./output/belief_accuracy_h2.png" width="90%" />
]

---
class:middle

### Effects on Polarization & Subjective Well-Being (H3 & H4)

.center[
<img src="./output/pol_sw_index.png" width="90%" />
]

---
class:middle

## Limitations

- Scope conditions: focus on **.red[direct effects]** of exposure to misinformation on WhatsApp.

- Effects might be **.red[context-dependent]**

- Previous deactivation studies found different results in United States (x2), France and Bosnia
  
- **.red[Election context]**: richer information environment + motivated reasoning

---
class:middle, center, inverse

# Tech & Public Policy Proposal

---
class:middle

## Multi-Country WhatsApp Deactivation

- Country Selection: Mexico, South Africa, India

- High WhatsApp usage + Ethnic and Political Polarization + Misinformation traveling primarely through WhatsApp

- Multimedia Deactivation - Focus on reducing exposure to Misinformation

- Reduction on WhatsApp usage - Full deactivation, >10min per day.

- Four week period.

---
class:middle

## Extensions: Experimental

- Non-Election Based Deactivation

- Effects of group-based conversations on political polarization (Fall 2024)

---

## Extensions: Data Donation on WhatsApp

.center[
<img src="monitor.png" width="50%" />
]

---
class:middle

## Progress

- Our first big effort was on building a team of experts and select the countries to run the experiment:

+ [Joshua Tucker](https://wp.nyu.edu/fas-joshuatucker/): Director of NYU’ Center for Social Media and Politics (CSMaP)

+ [Shelley Liu](https://www.shelleyxliu.com/): Assistant professor at Duke University, Sanford School of Public Policy

+ [Rajeshwari Majumdar](https://r-majumdar.github.io/): PhD Candidate, NYU Politics
    
  + [Carolina Torreblanco](https://carolina-torreblanca.github.io/): PhD Candidate, NYU Politics
  
  + [Maitreyi Natarajan]():  McCourt School’s Tech & Public Policy Scholar

---
class:middle

## Progress

- IRB Review: November and December

- Recruitment: Starting in January

- Deactivation: Early April to May.

---
class:middle, center, inverse

# Impact

---
class:middle

## Causal Effects of WhatsApp on Politics

--
- **Claims about WhatsApp**:

- Downstream effects on episodes offline violence 
  
  - Affects voting choices in Global South elections
  
  - Facilitates spread of misinformation -> beliefs

- **Our work provides the first comparative evidence for the effects of WhatsApp on political outcomes**.

- Separates general usage vs direct exposure to misinformation
  
  - Contextual variation
  
--

---
class:middle

## Causal Effects of Misinformation on Political Attitudes

- Since 2016, Google scholar returns 200,000 studies to the word "misinformation".

- **No causal evidence -> poor policy choices.**

- Policymakers and  technologists risk proposing solutions grounded in conventional wisdom, folk theory, or research from heterogeneous contexts. 
   
- Most recent studies (**.red[Meta 2020 Studies]**) have found null effects from misinformation exposure on attitudes.

---
class:middle

## The role of multimodal content on spread of misinformation

Our work also focuses on multimodal content:

- Most academic work has focuses on textual misinformation

- In contexts of low literacy, multimodal content have dominated the misinformation market

- travel fasters
  
  - easier to understand
  
- Most social media companies have guardrails focused on text/sources. Little developed for multimodal content.

---
class:middle

## Dissemination Plan
- Present at major academic conferences in the next year calendar: APSA, IC2S2 & ICA

- Summer 2024: Social Media and Politics Conference at Georgetown (together with John Ladd)

- Open discussions about industry-academic collaborations with Meta/WhatsApp

---
class:center, middle, inverse

# Q&A