KIBO NAGASAKI

he/him | age 16 | Missassauga, ON

Bronze Merit Award and Arcelor Mittal Dofasco Quality Systems Award, BASEF 2023

Edited by Cassidy Wideman

INTRODUCTION

In the 21st century, the economic growth in China, the collapse of the Arab Spring, and the rise of populism in European and South and North American politics have raised a question about the strength of democracy as a political system. This atmosphere was well described in “How Democracies Die” (Levitsky and Ziblatt 2018). The authors spent more than 20 years studying the breakdown of democracies in Europe and Latin America, and they demonstrated that our democracy is in danger. It has been a long time since some studies have argued that social network services (SNSs) are degrading democracy (e.g., Engesser et al. 2016, Tucker et al. 2019, Rau and Stier 2019, Miller and Vaccari 2020, Lewandowsky et al. 2020). SNSs, rather than public institutions or political parties, can facilitate change in society and democracy. Howard et al. (2011) concluded that SNSs played an important role in the Arab Spring. Sohn (2017) studied the distribution of Ice Bucket Challenge videos on SNSs and found that the videos were viewed by 440 million people worldwide, raising $220 million and positively impacting amyotrophic lateral sclerosis (ALS) research. This demonstrates how SNSs can quickly create new norms of social behaviour. Allcott and Gentzkow (2017), Shao (2017), and Vosoughi et al. (2018) analyzed fake and/or incorrect information during the 2016 U.S. presidential election and pointed out the important role of SNSs in the spread of fake and/or incorrect information. Recently, Lorenz-Spreen et al. (2022) carried out a systematic study providing causal and correlational evidence on the link between the global uptake of SNSs and the decline in democracy. Alongside the positive effects of SNSs on democracy, they showed evidence of serious threats to democracy and claimed an urgent necessity of research for a better understanding of the negative effects of SNSs on democracy.

When we consider the impact of SNSs on our society, \we have to consider the backfire effect — one manifestation of confirmation bias. The backfire effect is the tendency for people to resist accepting evidence/truth that conflicts with their beliefs. We also must pay attention to how much we trust governments, society, and related science and technology. One example is the various negative campaigns against the coronavirus (COVID-19) vaccines that have been deployed on SNSs. When discussing the impact of SNSs on democracy, the starting point is a case where the backfire effect can be ignored, as the effect of SNSs with the backfire effect can only be discussed through comparison. However, to my knowledge, there are no previous studies that have found that SNSs have a negative impact on democracy in cases where the backfire effect can be ignored.

On the other hand, SNSs are nothing more than information transmission technology. How the information conveyed by SNSs is expressed is also important. For example, the statements “90 out of 100 people died” and “10 out of 100 people survived” state the same information, but the impression of the information on recipients is different. Another example of the importance of how information is stated can be seen in Germany after World War I. The Weimar Constitution is the constitution of the German Republic established in 1919 after Germany's defeat in World War I and is an example of a modern democratic constitution. Considering that Germany was controlled by an imperial government before World War I and Hitler came to power on January 30, 1933, some think that Germany was less democratic, even after the World War I. However, the value of the democracy index, a measure of democracy maturity by the Varieties of Democracy (V-Dem) project, during 1920−1932 for Germany was lower than that of the UK, but higher than that of Canada and the USA (Varieties of Democracy 2022). Indeed, Hitler won power through democratic elections and became Chancellor. At that time, many people heard Hitler's message in person at lecture halls, over the radio, or in films. This implies that Hitler's speeches might have contained words and expressions that resonated with the German people after World War I but had a negative impact on democracy. However, to my knowledge, no study has comprehensively examined the words used in Hitler's speeches to determine whether Hitler's word choices influenced the deterioration of German democracy.

In this study, I hypothesized that the power of words and the method through which they are spread can influence democracy. To test this hypothesis, the following two studies were carried out. First, the relationship between the spread of information on Twitter about the shortage of toilet paper and actual toilet paper sales were studied. This was because the backfire effect and distrust of the government and technology seem to have little effect on people's buying of toilet paper. Secondly, the words Hitler frequently used in his speeches were analyzed. In this study, the words in Hitler's speeches from the German-Austrian bilateral Nazi party (National Socialist German Workers' Party) conference in Salzburg on August 7, 1920, to January 22, 1933, just before the Nazi party took power, were extracted and their characteristics were investigated.

MATERIALS & METHODS

Toilet paper shortage
In this study, the data on toilet paper sales and the spread of information on Twitter in Japan were examined. The daily sales record of toilet paper in Ontario or Canada could not be found, which can be freely accessed. On the other hand, the Japanese government collected and opened the sales records based on the point-of-sale (POS) data (Ministry of Economy, Trade and Industry 2023). Forty-five million Japanese individuals are using Twitter, and the spread of information can be easily traced. Furthermore, Twitter is the SNS with the highest rate of use to collect information in Japan (Ministry of Internal Affairs and Communications 2023).

The sales index is expressed as the change in sales compared to the previous year. In this study, the POS data of toilet paper from February 21, 2020, to March 10, 2020, was collected. This window of time was chosen as toilet paper sales in Japan peaked on February 28. The data on Twitter were collected using keywords such as “toilet paper” and “sold-out” in Japanese through search application programming interface (API). A total of 4,476,754 tweets were retrieved, 2,945,782 of which were retweets. The top 10,000 retweets of collected tweets were used in this study.

In this study, Twitter users were categorized into seven types as follows: (a) number of views of correction information only, (b) number of views of misinformation only, (c) number of views of sold-out information only, (d) number of views of both correction information and misinformation, (e) number of views of both correction information and sold-out information, (f) number of views of both misinformation and sold-out information, and (g) number of views of all of correction information, misinformation and sold-out information. To study which form of information spread led to the excessive purchasing of toilet paper and to avoid multicollinearity, principal component analysis and multiple regression analysis using principal components as variables were applied.

Frequently used words in Hitler’s speeches
In this study, the target period (August 7, 1920–January 22, 1933) was separated into two periods; the first half (August 7, 1920–March 18, 1930) and the second half (May 9, 1930–January 22, 1933). This was because on April 27, 1930, Goebbels was appointed Director of Public Relations for the Nazi Party. Goebbels was known as a “propaganda genius” who exerted great power in deifying Hitler and spreading Nazism through mass media, movies, and various events. On September 14, 1930, less than half a year after Goebbels took this position, the Nazis became the second largest party in the Reichstag. Therefore, in this study, the beginning date of the second period was set as May 9.

The contents of Hitler's speeches were extracted from books available at the McMaster University Library and books accessible online (Boepple 1925, Hitler 1939, Jäckel and Kuhn 1980, Saur 1992). The total numbers of speeches examined, and words extracted from them were 268 and 844,197, respectively. Frequently used words and terms frequently used together in Hitler’s speeches were analyzed using Sketch Engine.

RESULTS AND DISCUSSION

Toiler paper shortage
Figure 1 illustrates the change in the number of views of each categorized type during the target period. It was found that the numbers of type a (views of correction information only) and type e (views of both correction information and sold-out information) both peaked on February 28 which coincided with the peak of the sales index. The correlations among the seven explanatory variables (a to g) were very large.

Table 1 summarizes the eigenvectors and the contribution rate of each principal component. It was found that the 1st principal component could reproduce many of the original data due to the 0.885 contribution rate, but only one principal component model was not adequate to discuss the impact of SNSs. In this study, the 1st, 2nd, and 4th principal components were used to cover all explanatory variables (a to g).

**Table 1**: Eigenvectors and contribution rates of principal components.

Then, the multiple regression analysis was applied. However, since the principal component analysis was already applied, the 1st, 2nd, and 4th principal components were used as explanatory variables. The results of the multiple regression analysis using the 1st, 2nd, and 4th principal components are shown in Figure 2. The coefficient of determination was 0.980, suggesting that the regression results were accurate.

**Table 2**: Change in sales index from February 21 to March 10, 2020. Red line represents actual sales index and blue line multiple regression result.

Figure 2 Change in sales index from February 21 to March 10, 2020. Red line represents actual sales index and blue line multiple regression result.

The importance of each variable is summarized in Table 2. The values of variables (a) and (e) were the largest, meaning that the variables (a) and (e) had the most influence on the Sales Index. Namely, the number of views of correction information only (a) and the number of views of both correction information and sold-out information (e) are the most important. In other words, it was found that when the number of views of correction information only or that of both correction information and sold-out information increased on February 28, people were driven to hoard toilet paper, causing a shortage of toilet paper.

The results of this study indicate that the wide diffusion of correction information may have led to the social disruption of toilet paper hoarding in Japan. One possible explanation is pluralistic ignorance — a situation in which individuals mistakenly believe that everyone disagrees with them when in reality, most people share the same opinion. In the case of this study, the majority did not believe the misinformation about the shortage of toilet paper, but they considered that other people might believe the misinformation, anticipated that toilet paper would be sold out, and were then accelerated to purchase it for themselves.

**Table 2**: Importance of each variable.

The spread of misinformation and fake news can lead to social chaos. In particular, SNSs can easily spread information whose truth or facts are uncertain, causing social problems. It has been naively believed that correction information is effective in calming social panic caused by misinformation. However, the results of this analysis suggest that we must investigate the impact of SNSs on democracy deterioration further, including the spread of misinformation and fake news which accompanies the backfire effect, confirmation bias, and pluralistic ignorance.

Frequently used words in Hitler’s speeches
In Table 3, the nouns and the frequency with which they were used in Hitler’s speeches are summarized. These results are illustrated using word clouds in Figure 3. It was found that Hitler criticized the policies of “Stresemann”, who formed the German Folk Party in the Weimar Republic and served as Chancellor and Foreign Minister, and that “Marxism” and “Jews” were his targets of attack. On the other hand, it was found that Hitler mentioned "Jews" in his speeches predominantly in the first half of the period and decreased sharply in the second half. After winning the elections in 1930, anti-Semitism was withheld. This suggests that anti-Semitism was an ideology for unifying the party, not for taking Chancellor seat.

The words “France”, “Great Britain”, and “Italy”, the victorious nations of World War I, which were often mentioned in the first half, were not mentioned in the second half. In the second half, there were more references to “Papen”, who became Chancellor in place of Stresemann, who died in 1929 and worked hard to realize the Hitler Cabinet. Hitler criticized President Hindenburg for issuing “emergency decrees”. It is suggested that Hitler proceeded on the road to gaining power while referring to the “leadership”, the “constitution”, the “cabinet”, the “organization”, etc.

Until the Nazi party took over the core of the political world, by focusing on dissatisfaction with the victorious nations and anxiety of daily life of German people, Hitler attracted the attention of the people. From the stage when the government power came into view, Hitler criticized the Hindenburg and Weimar regimes which neglected the Diet. And in contrast, by touching on “farmers (namely citizens)” often, Hitler shifted the focus to how the Nazi Party promised the “future” of the German people.

The adjectives frequently used are shown in Figure 4. “Bavarian”, “national”, and “Munich”, which were characteristic words in the first half, were replaced with political and economic terms in the second half. This also indicates a change in the terminology used for gaining power. In addition, the representative adjective “international” in the first half, which was associated with capital, finance, Jews, etc., while the adjective “new” in the second half was associated with Germany, empire, era, etc. This is similar to the discussion with nouns; in the first half, the Nazi Party intended to instill their presence among the people by inciting public anxiety and dissatisfaction and strengthening the unity of the Nazi Party. In the second half, when the Nazi Party aimed to gain power, Hitler clarified what he promised to the people. Additionally, in the verb list (Figure 5), it was found that in the first half, a lot of unproductive verbs such as “die”, “protest” and “complain” were used, but in the second half, constructive verbs such as “overcome”, “fulfill” and “experience” were frequently used.

**Table 3**: Nouns frequently used in Hitler's speeches.

**Figure 3**: Word clouds of nouns frequently used in Hitler’s speeches. Words that were used more frequently are represented in larger text. A) August 7, 1920-March 18, 1930. B) May 9, 1930-January 22, 1930.

**Figure 4**: Word cloud of adjectives frequently used in Hitler’s speeches. Words that were used more frequently are represented in larger text. A) August 7, 1920-March 18, 1930. B) May 9, 1930-January 22, 1933.

**Figure 5**: Word clouds of verbs frequently used in Hitler’s speeches. Words that appear more frequently in each period are drawn larger. A) August 7, 1920 – March 18, 1930. B) May 9, 1930 – January 22, 1933.

Takada (2008) focused on the cuts of Hitler's speeches in the election campaign films after the seizure of power (March 1936 and March 1938) and clarified that these cuts had common linguistic features. In both of Hitler’s speeches, which lasted only a few seconds, Hitler used "Antitheses" which forced the audience to choose between black and white, and "Parallelisms" which imprinted a rhythmic impression on the audience. By repeating these two rhetorical methods, it was suggested that Hitler aimed to increase the inductive power and operability of his speeches to the audience. Hitler objectively analyzed what the people wanted to hear, and integrated the words and terms that best suited them with a way of speaking that was easy for the audience to hear, remember and support.

CONCLUSION

It is often said that the spread of false information, such as fake news, influences people's decisions and actions, and has a negative impact on democracy. However, in this study it was found that the spread of correction information had an impact on toilet paper hoarding behaviour in Japan. This finding suggests a need to identify the appropriate correction strategies based on the spread of both correction information and misinformation, to prevent the deterioration of democracy.

In Hitler's speeches, at first, he frequently used words that incited the dissatisfaction and anxiety of the German people after World War I and at the same time solidified the unity of the Nazi Party. After that, he frequently used words about what kind of promises he could make for the future of Germany. The words, expressions, and structure of a speech have the potential to resonate with people. Furthermore, there are effective measures to convey the speaker’s voice and gestures to people. However, my finding suggests that only when the people have a strong desire to hear it, its potential will become manifest, and the people can become enthusiastic about it even if it degrades democracy.

In the future, I will study what words, expressions, and discourse have been used in Hitler’s speeches and SNSs. Then, I will investigate how these have been combined with various information spread methods, such as SNSs, and how these have been able to influence people’s thoughts and decisions. Finally, I will elucidate how words, expressions, information spread methods, and desires of people are synchronized to determine how they can influence the deterioration of democracy and what measures can be used to prevent democracy from further deterioration.

ACKNOWLEDGEMENTS

I would like to thank BASEF (the Bay Area Science and Engineering Fair) for giving me the valuable opportunity to present this study. I thank my father for teaching me the German language.

REFERENCEs

Allcott, H., & Gentzkow, M. (2017). Social media and fake news in the 2016 election. Journal of Economic Perspectives, 31(2), 211–36. https://doi.org/10.1257/jep.31.2.211.

Boepple, E. (1925). Adolf Hitlers Reden. Mit seinem Bild von Otto v. Kursell, Munich, Germany, Deutscher Volksverlag.

Engesser, S., Ernst, N., Esser, F., & Büchel, F. (2016). Populism and social media: How politicians spread a fragmented ideology. Information, Communication and Society, 20(8), 1109 – 1126. https://doi.org/10.1080/1369118X.2016.1207697.

Hitler, A. (1939). Mein Kampf. London, the UK, Hurst and Blackett.

Howard, P.N., Duffy, A., Freelon, D., Hussain, M.M., Mari, W., & Maziad. M. (2011). Opening closed regimes: What was the role of social media during the Arab Spring? Retrieved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2595096.

Jäckel, E., & Kuhn, A. (1980). Hitler. Sämtliche Aufzeichnungen 1905-1924. Stuttgart, Germany, Deutsche Verlags-Anstalt.

Levitsky, S., & Ziblatt, D. (2018). How democracies die. New York, NY: Crown Publishing.

Lewandowsky, S., Smillie, L., Garcia, D., Hertwig, R., Weatherall, J. et al. (2020). Technology and democracy: Understanding the influence of online technologies on political behaviour and decision-making. JRC Science for Policy Report. https://doi.org/10.2760/709177.

Lorenz-Spreen, P., Oswald, L., Lewandowsky, S., & Hertwig, R. (2022). A systematic review of worldwide causal and correlational evidence on digital media and democracy. Nature Human Behaviour, 7, 74 – 101. https://doi.org/10.1038/s41562-022-01460-1.

Miller, M.L., & Vaccari, C. (2020). Digital threats to democracy: Comparative lessons and possible remedies. International Journal of Press-Politics, 25(3), 333 – 356. https://doi.org/10.1177/1940161220922323.

Ministry of Economy, Trade and Industry. (2023). BigData-STATS dashboard (β version). Retrieved from https://www.meti.go.jp/statistics/bigdata-statistics/bigdata_pj_2019/index.html (in Japanese).

Ministry of Internal Affairs and Communications. (2023). Survey report on COVID-19 information distribution, Retrieved from https://www.soumu.go.jp/main_sosiki/joho_tsusin/eng/pressrelease/ 2020/6/19_2.html.

Rau, J.P., & Stier, S. (2019). Die Echokammer-Hypothese: Fragmentierung der Öffentlishkeit und politishe Polarisierung durch digitale Medien? Zeitschrift für Vergleichende Politikwissenschaft, 13, 399 – 417 (in German).

Shao, C., Ciampaglia, G.L., Varol, O., Flammini, A., & Menczer, F. (2018). The spread of low-credibility content by social bots. arXiv preprint. Retrieved from
https://doi.org/10.48550/arXiv.1707.07592.

Sohn, E. (2017). Fundraising: The ice bucket challenges delivers. Nature, 550, S113 – S114. https://doi.org/10.1038/550S113a.

Takada, H. (2008). Hitler's Speeches in Election Campaign Films -Rhetoric of Contrast and Similarity-. Zeitschrift of Japanische Gesellschaft für Germanistik, 136, 113 – 130 (in Japanese).

Tucker, J.A., Guess, A., Barberá, P., Vaccari, C., Siegel, A., Sanovich, S., Stukal, D., & Nyhan, B. (2018). Social media, political polarization, and political disinformation: A review of the scientific literature. Retrieved from https://hewlett.org/wp-content/uploads/2018/03/Social-Media-Political-Polarization-and-Political-Disinformation-Literature-Review.pdf.

Varieties of Democracy. (2022). The V-Dem dataset. Retrieved from https://v-dem.net/data/the-v-dem-dataset/.

Vosoughi, S., Roy, D., & Aral, S. (2018). The spread of true and false news online. Science, 359(6380), 1146 – 1151. https://doi.org/10.1126/science.aap9559.

Deterioration of Democracy in the 21st century – Approach from toilet paper shortage through Twitter and speeches by Hitler