The January 6 Insurrection and America's Standing Abroad: Natural Experimental Evidence from Five Unexpectedly Interrupted Public Opinion Surveys

American geopolitical power partly relies on foreign public support for its leadership. Pundits worry that this support is evaporating now that the United States—which claims to be the world’s beacon of democracy—has itself experienced democratic back- sliding. I provide the first natural experimental test of this hypothesis by exploiting that the January 6 insurrection of the US Capitol unexpectedly occurred while Gallup was conducting nationally-representative surveys in India, Indonesia, Malaysia, Romania, and Vietnam. Because Gallup uses random digit dialing I can identify the effect by comparing US leadership approval among respondents that were interviewed just before, and just after, January 6, 2021. I find that the insurrection had no effect on US approval. If even a violent attempt to overturn a free and fair election does not affect US approval abroad it is unlikely that any other domestic anti-democratic event will.

American Voters Do Punish Overt Undemocratic Behavior at the Polls: Natural Experimental Evidence from the 2021 Insurrection of the U.S. Capitol

Existing research suggests that overt undemocratic behavior by elected officials is insufficiently punished by American voters to electorally discourage democratic backsliding. Evidence for this proposition comes primarily from hypothetical survey experiments with relatively weak treatments. I test this hypothesis using a natural experiment with a powerful treatment: Donald Trump's incitement of the insurrection of the U.S. Capitol on January 6, 2021. The insurrection was unexpected to the general public, did not coincide with other events that could plausibly affect public opinion, and occurred while Gallup was conducting a nationally representative survey using random digit dialing. Comparing vote choice intention among respondents that were interviewed just before, and just after, the insurrection occurred suggests that the insurrection caused a 10.8% decline in support for the Republican Party, and an 8.4% increase in support for the Democratic Party. Politicians interested in winning elections have strong incentives to avoid insurrection-like events from occurring.

Response trends in a national random digit dial survey

This paper reviews response trends over 24 consecutive quarters of a National Random Digit Dial telephone survey. Trends for response rates and refusal rates are studied as well as the components of response rate, namely, contact, cooperation and completion rates. In addition other rates, including answering machine, busy and no answer are studied. While refusal rates declined over the six year period, contact and cooperation rates significantly declined causing response rates to decline. Answering machine rates and busy rates also showed a significant increase over time. Finally, correlation’s among the variables of interest are presented. The response rate is negatively correlated with the busy rate, the answering machine rate and the no answer rate. Implications of the above trends are discussed.

„Warum wird der Notruf 112 gewählt?“ – Befragung zum Notrufverhalten der Berliner Bevölkerung

Zusammenfassung Hintergrund Steigende Notrufzahlen und vermehrte Inanspruchnahme notfallmedizinischer Versorgungsstrukturen sind derzeit Gegenstand umfassender gesundheitspolitischer Debatten in Deutschland. Die vorliegende Arbeit untersucht in einer repräsentativen Befragung die Kenntnisse, das Verhalten und mögliche Ursachen in Zusammenhang mit dem Notruf 112 im Land Berlin. Methode In Zusammenarbeit mit Infratest dimap wurde eine Telefonumfrage nach dem Random-digit-dialing(RDD)-Verfahren mit einer repräsentativen Zufallsstichprobe von 1002 Personen der deutschsprachigen Berliner Wohnbevölkerung zu Kenntnissen und dem Umgang mit dem Notruf 112 durchgeführt. Eingeschlossen wurden Menschen im Land Berlin, die zum Befragungszeitraum mindestens 14 Jahre alt waren und über einen Festnetzanschluss verfügten. In 12 Fragen wurden Informationen zum Wissensstand über den Notruf 112, alternative Versorgungsstrukturen, das Nutzungsverhalten der Befragten und alternative Hilfsangebote einschließlich standardisierter Fallbeispiele abgefragt. Experteninterviews wurden ebenfalls durchgeführt. Ergebnisse 58 % der Befragten gaben an, den Notruf 112 für medizinische Notfälle, 74 % für Notfälle der Brandbekämpfung zu kennen. 91 % der Befragten gaben an, dass sie den Notruf 112 nur bei lebensbedrohlichen medizinischen Situationen wählen würden. Befragte männlichen Geschlechts, ältere Menschen und Befragte mit einem niedrigeren Bildungsstand gaben häufiger an, den Notruf 112 auch bei nicht lebensbedrohlichen medizinischen Problemen in Anspruch zu nehmen. Alternativen zum Notruf kannten 56 % der Befragten, wobei insbesondere die hausärztliche Versorgung kaum eine Rolle spielte. Experten sehen die fehlende Verfügbarkeit von Alternativen als ursächlich für die zunehmende Frequenz von Notrufen. Schlussfolgerung Fehlende suffiziente Alternativen in der Notfallversorgung könnten bei der vermehrten Inanspruchnahme des Notrufs eine Rolle spielen. Dennoch scheint die Berliner Bevölkerung durchaus mit der richtigen Nutzung des Notrufs vertraut zu sein. Die Gesundheitskompetenz von vulnerablen Gruppen sollte in Bezug auf das Wissen zum Notruf 112 gestärkt werden. Eine Bevölkerungskampagne könnte einen sinnvollen Ansatz darstellen. Weiterhin muss die Verfügbarkeit alternativer Versorgungsformen verbessert werden.

Do American Voters Really Not Punish Overt Undemocratic Behavior at the Polls? Natural Experimental Evidence from the 2021 Insurrection of the U.S. Capitol

Existing research suggests that too few American voters hold politicians electorally accountable for overt undemocratic behavior to reasonably deter democratic backsliding. Evidence for this proposition comes primarily from hypothetical survey experiments with relatively modest treatments. I test this hypothesis using a natural experiment with a powerful real-world treatment: Donald Trump's incitement of the insurrection of the U.S. Capitol on January 6, 2021. The insurrection was unexpected to the general public, did not coincide with other events that could plausibly affect public opinion, and occurred while Gallup was conducting a nationally representative survey using random digit dialing. Comparing Republican Party support among respondents that were interviewed just before, and just after, the insurrection occurred suggests that the insurrection caused a 10.8% decline in support for the Republican Party. Voters predominantly moved to the Democratic Party, rather than Independent. Back-of-the-envelope calculations suggests that this electoral penalty is sufficient to decide presidential elections.

Do American Voters Really Not Punish Overt Undemocratic Behavior at the Polls? Natural Experimental Evidence from the 2021 Insurrection of the U.S. Capitol

Existing research suggests that too few American voters hold politicians electorally accountable for overt undemocratic behavior to reasonably deter democratic backsliding. Evidence for this proposition comes primarily from hypothetical survey experiments with relatively modest treatments. I test this hypothesis using a natural experiment with a powerful real-world treatment: Donald Trump's incitement of the insurrection of the U.S. Capitol on January 6, 2021. The insurrection was unexpected to the general public, did not coincide with other events that could plausibly affect public opinion, and occurred while Gallup was conducting a nationally representative survey using random digit dialing. Comparing Republican Party support among respondents that were interviewed just before, and just after, the insurrection occurred suggests that the insurrection caused a 10.8% decline in support for the Republican Party. Voters predominantly moved to the Democratic Party, rather than Independent. Back-of-the-envelope calculations suggests that this electoral penalty is sufficient to decide presidential elections.

Feasibility and Utility of Multi-Mode Data Collection in Random-Digit Dialing Cellphone Surveys For Young Adults: Evidence From a Survey Experiment (Preprint)

BACKGROUND Young adults’ early adoption of new cellphone technologies have created challenges to survey recruitment but offer opportunities to combine traditional random-digit dialing (RDD) sampling with web-mode data collection. The National Young Adult Health Survey (NYAHS) was designed to test the feasibility of this methodology. OBJECTIVE (1) Compare response rates across telephone- and web-modes, (2) assess sample representativeness, (3) examine the impact of sampling design and sample weighting on statistical precision, and (4) compare cigarette smoking prevalence to a gold standard national survey. METHODS The sampling frame was randomized to: (1) single-mode telephone interviews, (2) telephone to web sequential mixed mode, and (3) single-mode web survey, and 831 young adults (18-34 years) were recruited via RDD at baseline. A soft-launch was conducted prior to main-launch. We compared web-mode to pooled single-mode telephone and mixed-mode (telephone-mode) on the American Association for Public Opinion Research (AAPOR) response rate 3 (RR3) and for the screening (S-RR3) and extended surveys (E-RR3). Base-weighted demographic distributions were compared to the American Community Survey (ACS). Telephone- and web-mode samples were pooled and calibrated to the ACS to calculate design effects (DEFF) and compare cigarette smoking prevalence to the National Health Interview Survey (NHIS). Consistency of estimates was judged by overlap of 95% confidence intervals (CI). RESULTS RR3 was higher in telephone- than web-mode (24% and 30% vs. 6.1% and 12.5% for soft- and main-launch, respectively), which was reflected in S-RR3 and E-RR3. The telephone-mode over-represented men (57.1% vs. 50.9%) and those enrolled in college (40.3% vs. 23.8%) and under-represented those with a BA/BS+ (34.4% vs. 55%). The web-mode over-represented white, non-Latinos (70.7% vs. 54.4%) and those with some college (30.4% vs. 7.6%) and under-represented Latinos (13.6% vs. 20.7%) and those with a HS/GED (15.3% vs. 29.3%). The DEFF was 1.29 (subpopulation range: 0.96 to 1.90). The NYAHS cigarette smoking prevalence was consistent with NHIS overall (15%, CI 12.4% to 18% vs. 13.5%, CI 12.3% to 14.7%), with notable deviation among 18 to 24 year-olds (15.6%, CI 11.3% to 22.2% vs. 8.7%, CI 7.1% to 10.6%), and those with <BS/BA as highest level of education (24%, CI 19.3% to 29.4% vs. 17.1%, CI 15.6% to 18.7%). CONCLUSIONS RDD sampling for a web survey is not feasible for young adults due to its low response rate. However, combining this methodology with RDD telephone surveys may have great potential for including media and collecting auto-photographic data in population surveys