iiinova — where your digital self keeps evolving

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 27/01/2024

How @lydiamccomas.bsky.social looks at me while I tell her yet another idea for a paper that I’ll never write

0101

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 02/01/2024

Face-to-face conversations about divisive political issues between out-partisans can reduce affective polarization while conversations between in-partisans can cause more extreme political views, finds Fang et al. in an impressive field experiment in Germany www.econtribute.de/RePEc/ajk/aj...

How in-person conversations shape political polarization: Quasi-experimental evidence from a nationwide initiative

Figure 3: Effect of in-person conversations on ideological polarization

Figure 4: Effect on beliefs and attitudes toward people with opposing views

1105

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 23/12/2023

Check up on your friends who you didn't send Christmas cards: those who are depressed are less likely to send Christmas cards, especially Christians who are depressed, finds Gallagher et al. www.tandfonline.com/doi/full/10....

Christmas cards: are senders full of joy and good cheer?

Figure 1. Depression, frequency of sending Christmas cards of Christian groups.

Table 1. Sociodemographics and sending Christmas cards by depressive group

072

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 13/12/2023

Discussing politics in partisan echo chambers (i.e., homogenous groups) leads to affective polarization, compared to mixed discussion groups, finds @sarahobolt.bsky.social @katharinalawall.bsky.social & Tilley doi.org/10.1017/S000...

The Polarizing Effect of Partisan Echo Chambers

Figure 1. How Affective Polarization Is Correlated with Perceptions of Friends’ Voting Behavior

Note: 83% CI (thick), 95% CI (thin). Affective polarization is measured as the thermometer difference between parties.

Figure 4. The Effect of Group Composition on Affective Polarization

Note: Panel A: Mean affective polarization levels by experimental condition and time point. 83% CI (thick), 95% CI (thin). Panel B: Marginal effect of homogeneous versus mixed group on change in affective polarization. 95% CI. Excludes respondents who changed partisanship from W1 to W2 ( ). Full model results in Appendix 6 of the Supplementary Material.

1185

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 14/12/2023

Putin’s propaganda is effective: Leading up to the Ukrainian invasion, Putin's propaganda increased Russian support for military aggression against neighboring countries, finds Krishnarajan & Tolstrup. Great propaganda media effects study doi.org/10.1126/sciadv.adg1199

Pre-war experimental evidence that Putin’s propaganda elicited strong support for military invasion among Russians

Fig. 1. Provocations and Putin’s escalations increase support for military aggression among Russians.
The x axis presents the control condition and the three treatment conditions. The y axis denotes the estimated values of support for war (0 = oppose strongly, 1 = oppose somewhat, 2 = neither favor nor oppose, 3 = favor somewhat, and 4 = favor strongly) with 95% CIs. The left panel presents the main results of the full sample given by model 1. The right panel illustrates the main results across Putin supporters (dark blue) and opponents (light blue) given by model 2. The bars at the bottom of each graph illustrate the distribution in respondents’ answers on support for war across each treatment condition (green = oppose strongly, yellow = oppose somewhat, orange = neither favor nor oppose, dark orange = favor somewhat, and red = favor strongly). In the right panel, the bars only show distributions among Putin supporters.

Fig. 3. Security provocations and Putin’s escalation increase support for military aggression substantially.
Charts of the estimated distributions of support for war across security treatment conditions. Green equals opposition to war (oppose strongly or oppose somewhat), orange is undecided/indifferent (neither favor nor oppose), and red equals support for war (favor somewhat or favor strongly). The top row shows distributions across all respondents, the middle row across Putin supporters, and the bottom row across Putin opponents.

073

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 15/12/2023

Tolerance for Disagreement, as a personality trait, is a stronger predictor of affective polarization than partisanship and other personality traits, finds @luanarusso.bsky.social & Vanagt doi.org/10.31219/osf...

The impact of the personality trait “Tolerance for Disagreement” on Horizontal Affective Polarization

Figure 3: Beta scores for out-group dislike and social avoidance

13313

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 16/12/2023

Attention begets production: when social media content producers receive more engagement, they increase their production of content, finds Srinivasan combining observational and field experimental data karthikecon.github.io/karthiksrini...

Notes: This figure presents correlations between the attention that a Reddit post receives, as measured by the
number of comments, and various measures of content production by the post’s author over the next week. Each point
represents a one-comment bin, and bars represent 95% confidence intervals. The outcome in Plot A is Plog(score+1),
which is a quality-weighted measure of output. The outcome in Plot B is an indicator for if any posts are produced in
the next week, capturing the extensive margin. The outcome in Plot C is quantity, measured by the count of posts.
The outcome in Plot D is quality, measured by the average score of posts. All outcomes are demeaned by subreddit.

Notes: This figure repeats the difference-in-differences design of Figure 2 for two alternative outcomes. Plots A and
B consider the number of posts per day, an interpretable measure of the quantity of output. Viral producers post
0.068 more posts per day, which is 183% of the baseline of 0.037 posts per day. Plots C and D analyze effects on the
mean score conditional on posting, which is a measure of post quality. Going viral does not significantly change post
quality.

Notes: This figure replicates the difference-in-differences design of Figure 2 on TikTok. The outcome is a qualityweighted of output, where quality is measured by likes per view. Posts are viral if they surpass the 80th percentile of
the likes distribution. In Plot A, each point represents a 1 day bin. Event time 0 is the day that the viral or random
TikTok is created, and is excluded from the graph. Output increases by 0.049 units per day in the 30 days following
going viral TikTik relative to the random baseline, which is 279% increase over the pre-period rate of 0.017 units
per day. Plot B graphs heterogeneity in the treatment effect by the degree of virality. Each point is the output of
the difference-in-differences design estimated on the subset of posts that go viral within a two-percentile band of the
upvotes distribution. Posts in the first viral point received between 94-110 likes (80th-82nd percentile), while posts in
the tenth viral point received more than 3,583 upvotes (98

12810

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 17/12/2023

When people post misinformation and it's tagged as misinformation by a single person, the poster retreats into echo chambers, but when the post receives a collective tag (e.g., community notes), the poster is less likely to retreat into such an echo chamber, finds Kim et al., doi.org/10.48550/arX...

Fig 1. Misinformation Tagging and Outcomes Measurement. a, Individual misinformation tagging in
which individuals cite PolitiFact fact-checking articles. Collective misinformation tagging through
Community Notes platform, which selectively exposes verified misinformation tags that receive diverse
votes as helpful. b, Operationalization of tweet political and content diversity. Political diversity captures
whether a poster cites a source with opposing political stance (binary 0/1), assessed from the aggregate
stances of referenced sources. Content diversity captures whether a post discusses topics unfamiliar to the
author’s historical tweets (continuous), assessed with the distance between the poster’s average tweet and
a particular tweet within a contextual embedding (sentenceBERT pre-trained on Twitter)

Fig. 2. Political and Content Diversity Changing with the Intervention of Individual and Collective
Misinformation Tagging. a, Results from Interrupted Time Series (ITS) analysis. The x-axis denotes the
timeline of tweets posted before and after tagging, with negative values the number of weeks before
posting tagged tweets and positive values the number of weeks after. The y-axis denotes political and
content diversity, with dots capturing the average diversity score of the corresponding week, and error
bars indicating 95% confidence intervals. Solid lines connect the dots revealing trends of political and
content diversity before and after tagging, with gray dotted lines tracing the counterfactual trend if factchecks had not occurred. b, Illustration of political and content diversity dynamics before and after
tagging. Before individual and collective tagging, posters exhibit increased political and content diversity,
which increases the likelihood of encountering a fact-checker.

Fig. 3. Delayed Feedback (DF) Analysis. a, Pre- and post-treatment periods. Post-treatment (t1)
represents the time window when “treated” tweets are tagged but control tweets are not. Pre-treatment (t0)
represents the time window with equal duration of t1 when both treatment and control tweets remain
untagged. b, The effects of individual and collective misinformation tagging on political and content
diversity, which are estimated by the difference in pre-post changes (treatment - control) of the outcomes.

Fig. 4. Linguistic Characteristics of Fact-checking Messages. We present a univariate kernel density
function for continuous variables (toxicity, sentiment, length, delay) and a histogram for the categorical
variable (reading ease). The purple line represents the distribution within individual misinformation
tagging; the yellow line represents the distribution within collective tagging.

1288

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 27/11/2023

Although people enjoy shopping on Black Friday more, they find Cyber Monday to be more convenient and useful, with no differences between men and women, finds Swilley & Goldsmith doi.org/10.1016/j.jr...

Black Friday and Cyber Monday: Understanding consumer intentions on two major shopping days

Table 2 & 3. Black Friday and Cyber Monday statistics and CFA results

031

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 26/11/2023

Populist politicians are successful because voters agree with their policy positions, not because of their populist rhetoric, even when voters hold populist attitudes, finds Dai & Kustov. doi.org/10.1017/psrm...

The (in)effectiveness of populist rhetoric: a conjoint experiment of campaign messaging

Table 1. Effects of policy and populist rhetoric on vote choice

04813

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 24/11/2023

You'll undoubtedly see media reports of misbehavior on Black Friday but most shoppers are well behaved (thekeep.eiu.edu/fcs_fac/13 ) But, those that do misbehave, are those who expended the most effort in planning their shopping AND are highly impulsive (doi.org/10.1177/0887... )

An Analysis of Consumer Behavior on Black Friday

Table 1: Individual customer behavioral observations

A perfect storm for consumer misbehavior: shopping on Black Friday

Table 2. Principal Components Analyses: Related Statistics, Items, and Factor Loadings

081

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 30/10/2023

Anti-Muslim and anti-Jewish hate online stems from the same fringe, white supremacist communities who target both groups interchangeably, finds Hobbs, @nazita.bsky.social, Li & Lucas, suggesting we need to consider hate against both groups in tandem doi.org/10.1007/s111...

Anti-Muslim and anti-Jewish attitudes among respondents in the American Mosaic Project 2014 survey. Each jittered black point in the top panel of this figure represents the number of anti-Jewish and anti-Muslim attitudes reported by a White, Non-Hispanic respondent—the sum of yes responses to seven statements, including “They don’t share my morals or values” and “They want to take over our political institutions.” Purple ellipses represent a normal data ellipses for responses conditional on any problem (bottom left corner) and average conditional on a sum of problems across both groups greater than 10 (top right corner). Using the same data, the bottom two panels display the average number of problems attributed to a group conditional on the number of problems attributed to the other group

This figure displays change in mentions of Jews and Muslims or Arabs that contained predicted hate speech compared to January 2017 among Gab users who posted every month January 2017 through August 2018

The bottom panel above displays hate crimes and bias incidents recorded by ADL and CAIR by month. The top panels compares these two sources. Overall, hate crimes and bias incidents gradually decline after the 2016 election—until a very large spike in anti-Jewish incidents in late 2018. In addition to the overall patterns potentially related to the election and inauguration of Donald Trump, we also see a longer term shift in anti-Jewish hate crimes relative to anti-Muslim hate crimes in mid-2017. This second shift mirrors the activity on fringe social media sites around Unite the Right

Anti-Muslim or Arab and anti-Jewish aggravated assault, murders, manslaughter, arson, and kidnapping recorded in the FBI UCR data by month

9802410

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 06/11/2023

Excited to be presenting my research on misinformation exposure and effects tomorrow for Democracy Day at Stanford w/ @ryanmoore.bsky.social & Stanford Data Science

Is it fake? Understanding misinformation in politics.

0214

Reposted by Lydia M

Ross Dahlke @rossdahlke.bsky.social · 19/09/2023

So happy to be joining Stanford Data Science as a Data Science Scholar I will spend the next two years in this community continuing to build my data capture and digital ecosystem experimentation software and participating in the Stanford DS community datascience.stanford.edu/news/welcome...

datascience.stanford.edu

Welcome to the 2024-2025 cohort of Data Science Scholars

1191