Audience’s retention of the framing of media coverage of russian constitutional reform in 2020

('против', 0.042004384),

('цик', 0.03413156),

('клишас', 0.030988555),

('регион', 0.030741977),

('сенатор', 0.023595272)]),

The third topic revolves around postponing of the date of the vote due to external factors. It features such words as `vote', `rescheduling', `changing' and others that suggest a certain shift in timing, which can possibly relate to the coronavirus pandemic crisis of 2020 (1tv.ru, retrieved from https://www.1tv.ru/news/2020-03-29/382834-prezident_podpisal_ukaz_o_perenose_iz_za_koronavirusa_golosovaniya_po_popravkam_k_konstitutsii)

[('путин', 0.17351648),

('голосование', 0.1484788),

('изменение', 0.073871315),

('рассказать', 0.0358519),

('дата', 0.028427312),

('сообщить', 0.027037216),

('назвать', 0.020420942),

('допустить', 0.019838767),

('вопрос', 0.01962853),

('перенос', 0.017952971)]),

The fourth topic suggests that it highlights the stage of announcement of the reform, with such words as `suggest', `constitutional', `referendum' and such. However, this topic possesses a significant difference from all the others. Not only does it contain the information about the procedure and changes considering it, it, unlike others, highlights the contents of one of the reforms. It is the value-oriented one regarding the religious aspect of the cultural and historical continuity in Russia (RBC, retrieved from https://www.rbc.ru/politics/02/03/2020/5e5d010e9a794718c0f57783) However, it does not hold as much presence as all of the other words related to a different topic. Therefore, it can be concluded that the main theme made salient here is the one of the initiation of the reform, and also of its approval by the Constitutional Court of Russian Federation, due to the fact that it features both terms: `court' and `constitutional'.

[('предложить', 0.10707044),

('госдума', 0.0849836),

('конституционный', 0.07309438),

('послание', 0.033780668),

('суд', 0.028216813),

('бог', 0.026424328),

('референдум', 0.026408633),

('закон', 0.025360083),

('принять', 0.023175169),

('совет', 0.02250217)])]

8). The visualization of the results has been enacted with the usage of the pyLDAvis library, which helps visualize the set of topics and the top of the most salient words.

The graph created with pyLDAvis can be found in Appendix 1.

Identifying Framing Bias: Results

As it has been mentioned previously, we have been able to identify 4 distinctive topics regarding the constitutional reform in Russia. Two of them highlight the initial stage of the reform, including its proposition which coincided with the resignation of Russian government. One of those topics that did not highlight the resignation, highlighted the initiative itself, as well as its approval by the Constitutional Court of Russian Federation.

Another topic refers to the shifts in the scheduling of the popular vote. And the last one highlights the activity of the task force assembled to work on the new constitutional reform.

However, the topics discussing the contents of the reform are by far less present in the article headlines we have analyzed. All of the aforementioned topics deal with procedural details and rescheduling changes but not with the nature of the reform itself. Thus, we can suggest that procedural complications and updates was an aspect made more salient in online media at the moment the data was collected. The frame with which the news about the constitutional reform are presented is likely to be the one built on paying attention to the process of the reform rather than its contents.

Identifying Selective Exposure: Design and Data

Previous studies have shown that selective exposure can be evoked by a shift in preferences so vast and influential it suggests a choice of an entire medium over the other. For example, a certain predisposition may lead a person to engage more with television programs than with printed press. Such fluctuations are known to affect making of judgements as well as decision-making, especially when it comes to political decisions (Sweeney & Gruber, 1984).

To analyze the relations between selective exposure and judgement making, we are going to be using the data collected by WCIOM in regards with the constitutional reform in Russia and the judgements people make about its main incentives (WCIOM, retrieved from https://wciom.ru/index.php?id=236&uid=10220).

The survey was held from 29th until 31st of March, 2020 and it involved 1600 Russians above the age of 18. This telephone-based survey was developed with two stage stratified random sampling. The phone numbers have been retrieved from the exhaustive list of functional Russian phone numbers. The data has been weighted in accordance with demographical parameters and probability. The maximum error is not more than 2,5% with 95% confidence. The choice of data provided by this survey was driven by its recognizable external validity.

Moreover, previous studies have shown that unstable in its effects selective exposure bias gains strength if it is connected to personal issues, attitudes and those aspects of life that evoke a stronger emotional response (Taber & Lodge, 2006). The questions of which the survey is comprised of deal with such personal questions and attitudes and not with the actual socio-political aspect of the new reform. Thus, it is more suitable for our research.

The first part of the questionnaire is comprised of questions about the socio-demographic profile of a respondent, including those about their age, sex, level of education and the region of origin.

The next block includes a set of constitutional reform initiatives, all of which respondents can assess with one of the following statements: `approve', `do not approve', `cannot make up my mind'.

The questionnaire only considers a very specific set of largely populist initiatives related to attitudes rather than politics and makes up a list of the following initiatives each respondent has to form a judgement about (the summary of each initiative is provided down below):

- To solidify the sovereignty of the Russian Federation in the Constitution

- To forbid civil servants to hold dual citizenship or be in possession of any assets outside of the Russian Federation

- For Russia to take action in maintaining the world peace and security as well as peaceful coexistence of states and nations

- The primacy of the Constitution of the Russian Federation over the international law

- To secure the status of the institution of marriage as the one of a union of a man and a woman

- The safeguarding of historical truth

- Recognition of the Russian Federation as the successor of the Soviet Union on its own territory

- Recognition of the Russian language as the language of the constituent people

- Recognition of culture in the Russian Federation as a unique legacy of its multinational people

If the questions of this block inquire whether the respondents consider these initiatives to be important or not, the next one, presenting the same set of initiatives, asks whether or not the respondents stand in support for implementing these changes into the Constitution.

The following block deals largely with media consumption. Respondents have been asked whether or not they watch television, and if so, how often they engage in TV watching. The same question has been asked about respondents' engagement with the Internet in the very same manner.

Two of the following questions were the ones regarding the respondents' financial situation as well as their own subjective evaluation of it, ranging from `very good' to `very bad'.

These questions were followed up by a couple of those regarding people's occupation, as well as a question about whether a respondent is considered to be a state employee or not.

Identifying Selective Exposure: Analysis

WCIOM has also evaluated all of the acquired data. According to WCIOM, the most important initiative is considered to be the one regarding the safeguarding of historical truth. The number of people in favor of this view comprises of 89%. The amount of people who stand in support of this change amounts to 90%.

The initiative about recognizing the culture of the Russian Federation as a unique legacy of its multinational people comes in second, with every tenth person disagreeing with its importance. However, 87% still seem to be in favor of it.

86% of Russians consider the recognition of the Russian language as the language of the constituent people to be an important addition to the Constitution. 11% of respondents have objected to it.

83% of people stand in favor of securing the status of the institution of marriage as the one of a union of a man and a woman. 15% of people have considered such a change to be of no importance to them.

The last of the most popular ones is the initiative regarding recognition of the Russian Federation as the successor of the Soviet Union on its own territory. 81% of respondents have evaluated it as an important addition to the current Constitution, 14% of people have disagreed with that statement (WCIOM, retrieved from https://wciom.ru/index.php?id=236&uid=10220).

However, what is of most interest to us is the previously formulated hypothesis.

H2. Audiences' selective exposure is connected to their judgements regarding political decision-making.

H0. Audiences' selective exposure is connected to their judgements regarding political decision-making.

As have been stated before, topics related to personal issues and attitudes are more likely to evoke selective exposure (Taber & Lodge, 2006). All of the questions concerning the new constitutional initiatives are the ones related to attitudes and not much else.

There is also evidence that people tend to develop preferences for certain media formats when the selective exposure effects are especially strong. This could make a person opt for only television programs or only online media, only printed press etc. (Sweeney & Gruber, 1984).

The current survey provides us with two forms of media consumption: preferably television or preferably the Internet. Thus, we would like to see if there is indeed connection between exhibiting certain media preferences as a consequence of selective exposure bias and the judgement making process when it comes to evaluating the initiatives if the new Constitutional reform.

To test this hypothesis, we apply the optimal fit module from R to the dataset provided by WCIOM. The detailed result is a part of Appendix 2. A regression model explaining around 40% of all instances was built.

However, the correlation between the media consumption habits and evaluation of statements has not been spotted. Thus, confirming the H0 hypothesis and disproving the hypothesis H2.

Identifying Selective Exposure: Results

A connection that has been identified however is the one between the parameter of age and assessing the given statements. As it turns out, the age factor is stronger when it comes to forming an opinion about the new Constitutional reform initiatives, and the media exposure habits are not that significant.

Nevertheless, a correlation between the media consumption that mostly leans towards television watching and the age has also proven to be present. Thus, suggesting that the age factor may influence not only judgement making, but also media consumption itself, even more so than political leanings and previously held beliefs do. However, these findings only apply to this one very unique situation of respondents' assessing the initiatives that are strongly attitudinal.

Overall Results and Discussion

The initial goal of this research lied in identifying frames that constitute media bias in online media coverage of the constitutional reform in Russia. However, we also aimed to find out whether media audience's selective exposure, driven by previously held beliefs and attitudes affects their retention of information and thus the way they evaluate the issues at hand. Identifying such an effect would be fruitful due to the fact that both framing media bias and selective exposure bias tend to influence readers' retention of information and therefore their political decision-making.

As for the first part, we were indeed able to identify a set of predominant topics regarding the constitutional reform. However, all of them were centered predominantly around the procedural aspect of the issue. According to theory, framing bias is an instance of making one aspect of the complicated topic more salient and visible than the others (Goffman, 1974). In this case, all of the media outlets we have analyzed tend to exhibit a form of framing bias where they shift their focus to the procedural aspects.

We have been able to identify four of the following topics. One of them points at the first proposition of the Constitutional reform, along with the resignation of Russian government in early 2020. Another topic brought to light not only the reform itself and its initiation, but also its approval by the Constitutional Court of Russian Federation. Rescheduling of the day of the popular vote amidst a crisis was another significant topic. The fourth topic highlighted the activity of the task force assigned to work on the contents of the Constitutional reform.

However, not much attention was paid to the contents of the reform itself. Thus, we conclude that the topics regarding the actual text of the reform may have been less salient amidst those highlighting the procedural aspect of the process.

Nevertheless, we managed to get some closure regarding the evaluation of the new initiatives when we were working on our second hypothesis. The data has shown that many people do in fact tend to support the attitudinal initiatives presented in the text of the reform.

As for the correlation between selective exposure bias and the judgement making regarding the attitudinal initiatives of the constitutional reform, no such correlation has been found. However, it has been noted that judgement making does in fact connect to the age factor of the respondents.

Delimitations and Suggestions for Future Research

Due to the fact that the data with such external validity have only included a number of attitudinal points to evaluate, it would be of interest to conduct a similar research regarding other initiatives from the project of the Constitutional Reform.

Moreover, the connection between the evaluation of those points and the age of the respondents suggests many possibilities for future research, regarding the exploration of this effect.

Judging by the way the effect of selective exposure bias was overestimated in this iteration of research, exploring the effects that in turn influence selective exposure in the context of gaining knowledge about the reform could be another venue in these studies.

CONCLUSION

The overall goal of this research was to identify framing bias in online media coverage of the constitutional reform in Russia.

The objectives were to develop the theoretical framework for the study of framing bias in online media; to explore the nature of the coverage of the constitutional reform; to note if exposure tendencies relate to audiences' retention-based judgements and decisions.

The problem of the research at hand was considered to be a tendency of both framing bias and selective exposure bias to influence readers' retention of information and therefore judgement- and decision-making.

The research question inquired whether the online media coverage of the constitutional reform in indeed features instances of framing bias, as well as whether media audiences' succumb to selective exposure bias enough for it to influence their decisions and judgements.

The following hypotheses have been formulated.

H1. Framing bias is present in the coverage of the constitutional reform in Russia.

H2. Audiences' selective exposure is connected to their judgements regarding political decision-making.

However, the second one did not end up being supported due to the fact that the connection between the evaluations of the reform point presented to the audience did not correlate with their overall media preferences.

The process of working on the first hypothesis in fact yielded results as we were able to identify several more salient topics in the stream of news messages about the constitutional reform.

Moreover, all of the topics tend to have an underlying similarity of them highlighting only one aspect of an issue, procedural complications, thus, we can suggest that the instance of framing media bias indeed occurred.

The research at hand contributes to the studies of frame media bias, exploring an instance of not politically charged frame in relation to popular vote.

Bibliography

1). American Association of Newspaper Editors, 1999. Examining our Credibility: Examining Credibility, Explaining Ourselves. Reston, Va. (http://www.asne.org).

2). Ansolabehere, Stephen, Snowberg, Erik C., Snyder Jr., James M., 2004. Unrepresentative information: the case of newspaper reporting on campaign finance. Working Paper. MIT

3). Atre, J., & Katz, E. (2005). What's killing television news? Experimentally assessing the effects of multiple channels on media choice. Paper presented at the International Communication Association Conference, New York. NY.

4). Best, S. J., Chmielewski, B., & Krueger, B. S. (2005). Selective exposure to online foreign news during the conflict with Iraq. Harvard International Journal of Press/Politics, 10(4).

5). Blei, D. M., Ng, A. Y., Jordan, M. I. (2003). Latent Dirichlet Allocation. Journal of Machine Learning Research 3, 993-1022.

6). Bovitz, Gregory L., Druckman, James N., Lupia, Arthur, 2002. When can a news organization lead public

7). Brockner, J., & Rubin, J. Z. (1985). Entrapment in escalating conflicts--a social psychological analysis. New York: Springer.

8). Budak, C., Goel, S., Rao, J. M., & Zervas, G. (2016). Understanding Emerging Threats to Online Advertising. Boston U. School of Management Research Paper No. 2505643.

9). Chaffee, S. H., & Miyo, Y. (1983). Selective exposure and the reinforcement hypothesis: An intergenerational panel study of the 1980 presidential campaign. Communication Research, 10, 3-36.

10). Chaffee, S. H., Saphir, M. N., Graf, J., Sandvig, C, & Hahn, K. S. (2001). Attention to counter-attitudinal messages in a state election campaign. Political Communication, 18(3).

11). Chong, Dennis & Druckman, James. (2007). Framing Theory. Annual Review of Political Science. 10. 10.1146/annurev.polisci.10.072805.103054.

12). Cotton, J. L. (1985). Cognitive dissonance in selective exposure. In D. Zillman & J. Bryant (Eds.), Selective exposure to communication (pp. 11-33). Hillsdale, NJ: Erlbaum.

13). Donsbach, W. (1991). Exposure to political content in newspapers: The impact of cognitive dissonance on readers' selectivity. European Journal of Communication, 6(2).

14). Edy, Jill & Meirick, Patrick. (2007). Wanted, Dead or Alive: Media Frames, Frame Adoption, and Support for the War in Afghanistan. Journal of Communication. 57. 119 - 141.

15). Entman, R. M. (1993). Framing: Toward clarification of a fractured paradigm. Journal of Communication, 43(4), 51-58.

16). Eveland, W. P. (2004). The effects of political discussion in producing informed citizens: The roles of information, motivation, and elaboration. Political Communication, 21(2), 177-193.

17). Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA: Stanford University Press.

18). Galston, W. A. (2003). If political fragmentation is the problem, is the Internet the solution? In D. M. Anderson & M. Cornfield (Eds.), The civic web: Online politics and democratic values. Oxford: Rowman & Littlefield Publishers. 35-144.

19). Gamson, W. A., & Modigliani, A. (1987). The changing culture of af?rmative action. Research in Political Sociology, 3, 137-177.

20). Gans, H. (1979). Deciding what's news. New York: Pantheon Books.

21). Geller, E. S., & Pitz, G. F. (1968). Confidence and decision speed in the revision of opinion. Organizational Behavior & Human Performance, 3(2), 190-201.

22). Gentzkow, M., Shapiro J. M. (2010). What Drives Media Slant? Evidence from U.S. Daily Newspapers. Econometrica, 78(1), 35-71.

23). Gilens, M., Hertzman, C. (2000). Corporate Ownership and News Bias: Newspaper Coverage of the1996 Telecommunications Bill. Journal of Politics, 62(2), 369-386.

24). Godden, D. (1976). Transition structure versus commitment in sequential subjective probability revision. Acta Psychologica, 40, 21-28.

25). Goffman, E. (1974). Frame analysis: An essay on the organization of experience. Harvard University Press.

26). Green, D., Palmquist, B., & Schickler, E. (2002). Partisan hearts & minds: Political parties and the social identities of voters. New Haven: Yale University Press.

27). Groseclose T., & Milyo J. (2005). A Measure of Media Bias. The Quarterly Journal of Economics, 120(4), 1191-1237.

28). Groseclose, Tim, Milyo, Jeff, 2004 (September). A Measure of Media Bias. Working Paper. UCLA.

29). Hamilton, James T., 2004. All the News That's Fit to Sell: How the Market Transforms Information into News.

30). Iyengar, S., & Hahn, K. S. (2009). Red media, blue media: Evidence of ideological selectivity in media use. Journal of Communication, 59(1), 19-39.

31). Iyengar, S, & McGrady, J. (2007).Mediaandpolitics: A citizen's guide. NewYork:W.W.Norton

32). Johnston, L. (1996). Resisting change: Information-seeking and stereotype change. European Journal of Social Psychology, 26, 799-825

33). Jonas, E., Schulz-Hardt, S., & Frey, D. (2005). Giving advice or making decisions in someone else's place: The influence of impression, defense, and accuracy motivation on the search for new information. Personality and Social Psychology Bulletin, 31(1), 977-990.

34). Jonas, E., Schulz-Hardt, S., Frey, D., & Thelen, N. (2001). Confirmation bias in sequential information search after preliminary decisions: An expansion of dissonance theoretical research on selective exposure to information. Journal of Personality and Social Psychology, 80(4), 557-571.

35). Kahneman, D., & Tversky, A. (1984). Choices, values, and frames. American Psychologist, 39(4), 341-350.

36). Kinder, D. R. (2003). Communication and politics in the age of information. In D. O. Sears, L. Huddy, & R. Jervis (Eds.), Oxford handbook of political psychology (pp. 357-393). Oxford: Oxford University Press

37). Klapper, J. T. (1964). The effects of mass communication. New York: Free Press.

38). Kull, S., Ramsay, C, & Lewis, E. (2003-4). Misperceptions, the media, and the Iraq war. Political Science Quarterly, 118(4), 569-598.

39). Li Susan. Latent Semantic Analysis & Sentiment Classification with Python. (2018. September 18). Retrieved February 20, 2020, from https://towardsdatascience.com/latent-semantic-analysis-sentiment-classification-with-python-5f657346f6a3.

40). Lott, J. & Hassett, K. (2004). Is Newspaper Coverage of Economic Events Politically Biased?. Public Choice, 160(1-2).

41). Marcus, G. E., Neuman, W. R., & MacKuen, M. (2000). Affective intelligence and political judgment. Chicago: University of Chicago Press.

42). McQuail, D. (2005). Mass communication theory (5th ed.). London: Sage

43). Meffert, M. R, Chung, S., Joiner, A. J., Waks, L., & Garst, J. (2006). The effects of negativity and motivated information processing during a political campaign. Journal of Communication, 56(1), 27-51.

44). Morstatter F., Liang Wu, Yavanoglu U., Corman S. R., & Huan Liu. (2018). Identifying Framing Bias in Online News. ACM Trans. Soc. Comput. 1(2), 5-17.

45). Mullainathan, Sendhil and Shleifer, Andrei, 2003. The Market for News.Q Working paper, MIT and Harvard

46). Mutz, D. C, & Martin, P. S. (2001). Facilitating communication across lines of political difference: The role of mass media. American Political Science Review, 95(1), 97-1 14.

47). Pan, Z., & Kosicki, G. M. (1993). Framing analysis: An approach to news discourse. Political Communication, 10(1), 55-75.

48). Pedro Henrique Calais Guerra, Adriano Veloso, Wagner Meira, Jr., and Virgнlio Almeida. 2011. From bias to opinion: A transfer-learning approach to real-time sentiment analysis. In Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD'11). ACM, New York, 150-158.

49). Pfau, M. J., Houston, B., & Semmler, S. M. (2007). Mediating the vote: The changing media landscape in U.S. presidential campaigns. Lanham, MD: Rowman and Little?eld.

50). Pinkley, R. L., Griffith, T. L., & Northcraft, G. B. (1995). "Fixed pie" a la mode: Information availability, information processing, and the negotiation of suboptimal agreements. Organizational Behavior and Human Decision Processes, 62, 101-112.

51). Price, V., & Tewksbury, D. (1997). News values and public opinion: A theoretical account of media priming and framing. In G. A. Barett & F. J. Boster (Eds.), Progress in communication sciences: Advances in persuasion (Vol. 13, pp. 173-212). Greenwich, CT: Ablex.

52). Puglisi, R. (2011). Being the New York Times: the political behaviour of a newspaper. The B.E. Journal of Economic Analysis & Policy,11(1), 20.

53). Puglisi, Ricardo, (2004). Being the New York Times: the Political Behavior of a Newspaper. Working Paper.

54). Reddy, R. R., Dugenpudi, S. R., Mamidi, R. (2019). Proceedings of the Second BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP. Association for Computational Linguistics, 77-84.

55). Redlawsk, D. P. (2002). Hot cognition or cool consideration? Journal of Politics, 64(4), 102

56). Scheufele, D. A. (1999). Framing as a theory of media effects. Journal of Communication, 49(1), 103-122.

57). Scheufele, D. A., & Tewksbury, D. (2007). Framing, agenda setting, and priming: The evolution of three media effects models. Journal of Communication, 57(1), 9-20.

58). Schulz-Hardt, S. (1997). Realitdtsflucht in Entscheidungsprozessen: Von Groupthink zum Entscheidungsautismus [Loosing reality in decision processes: From groupthink to autistic decision making], Bern, Switzerland: Huber.

59). Schulz-Hardt, S., Frey, D., Luthgens, C, & Moscovici, S. (2000). Biased information search in group decision making. Journal of Personality and Social Psychology, 78, 655-669.

60). Sean A. Munson and Paul Resnick. 2010. Presenting diverse political opinions: How and how much. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. ACM, 1457-1466.

61). Sears, D. O. (1968). The paradox of de facto selective exposure without preference for supportive information. In R. P. Abelson, E. Aronson, W. J. McGuire, T. M. Newcomb,

62). Sears, D. O., & Freedman, J. F. (1967). Selective exposure to information: A critical review. Public Opinion Quarterly, 31, 194-213.

63). Shoemaker, P. J., & Reese, S. D. (1996). Mediating the message: Theories of influences on mass media content. White Plains, N.Y: Longman.

64). Shoemaker, Pamela & Vos, Tim. (2009). Gatekeeping theory. Taylor & Francis.

65). Smith, T. J., III, Lichter, S. R., & Harris, T. (1997). What the people want from the press. Washington, DC: Center for Media and Public Affairs.

66). Stromberg, David, 2004. Mass media competition, political competition and public policy. Review of Economic Studies 71, 265 - 284

67). Sweeney, P. D., & Gruber, K. L. (1984). Selective exposure: Voter information preferences and the Watergate affair. Journal of Personality and Social Psychology, 46, 1208-1221.

68). Taber, C. S., & Lodge, M. (2006). Motivated skepticism in the evaluation of political beliefs. American Journal of Political Science, 50(3), 755

69). Tversky A., and Kahneman D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453-458.

70). Valentino, N. A., Hutchings, V. L., Banks, A. J., & Davis, A. K. (2007). Selective exposure in the Internet age: Emotional triggers of political information seeking. Paper presented at the American Political Science Association Conference, Chicago, IL

71). Vertzberger, Y. Y. I. (1990). The world in their minds: Information processing, cognition, and perception in foreign policy decisionmaking. Stanford, CA: Stanford University Press.

72). Zaller, J. (1992). The nature and origin of mass opinion. New York: Cambridge University Press.

APPENDIX

Call:

lm(formula = AGE ~ ., data = df_numeric)

Residuals:

Min 1Q Median 3Q Max

-30.4991 -6.2443 0.4467 6.4085 26.5503

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 62.5443887 2.3854054 26.220 <2e-16 ***

EDU 0.8547520 0.3840855 2.225 0.0263 *

FO -0.0855667 0.1431664 -0.598 0.5502

TIP 0.0672553 0.1686548 0.399 0.6902

b2_1 -0.0140207 0.0230863 -0.607 0.5438

b2_2 -0.0072580 0.0242219 -0.300 0.7645

b2_7 0.0266657 0.0193665 1.377 0.1689

b2_11 0.0001635 0.0137309 0.012 0.9905

b2_13 -0.0235033 0.0235302 -0.999 0.3181

b2_16 -0.0358851 0.0237645 -1.510 0.1314

b2_21 -0.0251689 0.0170312 -1.478 0.1398

b2_22 0.0045510 0.0255501 0.178 0.8587

b2_23 0.0004607 0.0289876 0.016 0.9873

b3_1 -0.0481448 0.0270551 -1.780 0.0755 .

b3_2 -0.0081128 0.0303446 -0.267 0.7893

b3_7 0.0025536 0.0261666 0.098 0.9223

b3_11 0.0114517 0.0159150 0.720 0.4720

b3_13 0.0136633 0.0285538 0.479 0.6324

b3_16 0.0422981 0.0362655 1.166 0.2438

b3_21 -0.0083418 0.0195494 -0.427 0.6697

b3_22 -0.0068688 0.0276356 -0.249 0.8038

b3_23 0.0296390 0.0306906 0.966 0.3344

d1 0.0941627 0.0182362 5.164 3e-07 ***

TV -2.4246656 0.2172637 -11.160 <2e-16 ***

DOHOD_0 -0.0309518 0.1035394 -0.299 0.7651

DOHOD -0.0433732 0.0465926 -0.931 0.3522

PROF -2.8426620 0.1709748 -16.626 <2e-16 ***

PROF1 0.0155224 0.0200002 0.776 0.4379

d3 -0.0091206 0.0138903 -0.657 0.5116

Signif. codes: 0 `***' 0.001 `**' 0.01 `*' 0.05 `.' 0.1 ` ' 1

Residual standard error: 9.71 on 874 degrees of freedom

(697 observations deleted due to missingness)

Multiple R-squared: 0.4137, Adjusted R-squared: 0.3949

F-statistic: 22.02 on 28 and 874 DF, p-value: < 2.2e-16

>

> optimal_fit <- step(fit_full, direction = 'forward')

Start: AIC=4133.91

AGE ~ EDU + FO + TIP + b2_1 + b2_2 + b2_7 + b2_11 + b2_13 + b2_16 +

b2_21 + b2_22 + b2_23 + b3_1 + b3_2 + b3_7 + b3_11 + b3_13 +

b3_16 + b3_21 + b3_22 + b3_23 + d1 + TV + DOHOD_0 + DOHOD +

PROF + PROF1 + d3

> summary(optimal_fit)

Call:

lm(formula = AGE ~ EDU + FO + TIP + b2_1 + b2_2 + b2_7 + b2_11 +

b2_13 + b2_16 + b2_21 + b2_22 + b2_23 + b3_1 + b3_2 + b3_7 +

b3_11 + b3_13 + b3_16 + b3_21 + b3_22 + b3_23 + d1 + TV +

DOHOD_0 + DOHOD + PROF + PROF1 + d3, data = df_numeric)

Residuals:

Min 1Q Median 3Q Max

-30.4991 -6.2443 0.4467 6.4085 26.5503

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 62.5443887 2.3854054 26.220 <2e-16 ***

EDU 0.8547520 0.3840855 2.225 0.0263 *

FO -0.0855667 0.1431664 -0.598 0.5502

TIP 0.0672553 0.1686548 0.399 0.6902

b2_1 -0.0140207 0.0230863 -0.607 0.5438

b2_2 -0.0072580 0.0242219 -0.300 0.7645

b2_7 0.0266657 0.0193665 1.377 0.1689

b2_11 0.0001635 0.0137309 0.012 0.9905

b2_13 -0.0235033 0.0235302 -0.999 0.3181

b2_16 -0.0358851 0.0237645 -1.510 0.1314

b2_21 -0.0251689 0.0170312 -1.478 0.1398

b2_22 0.0045510 0.0255501 0.178 0.8587

b2_23 0.0004607 0.0289876 0.016 0.9873

b3_1 -0.0481448 0.0270551 -1.780 0.0755 .

b3_2 -0.0081128 0.0303446 -0.267 0.7893

b3_7 0.0025536 0.0261666 0.098 0.9223

b3_11 0.0114517 0.0159150 0.720 0.4720

b3_13 0.0136633 0.0285538 0.479 0.6324

b3_16 0.0422981 0.0362655 1.166 0.2438

b3_21 -0.0083418 0.0195494 -0.427 0.6697

b3_22 -0.0068688 0.0276356 -0.249 0.8038

b3_23 0.0296390 0.0306906 0.966 0.3344

d1 0.0941627 0.0182362 5.164 3e-07 ***

TV -2.4246656 0.2172637 -11.160 <2e-16 ***

DOHOD_0 -0.0309518 0.1035394 -0.299 0.7651

DOHOD -0.0433732 0.0465926 -0.931 0.3522

PROF -2.8426620 0.1709748 -16.626 <2e-16 ***

PROF1 0.0155224 0.0200002 0.776 0.4379

d3 -0.0091206 0.0138903 -0.657 0.5116

Signif. codes: 0 `***' 0.001 `**' 0.01 `*' 0.05 `.' 0.1 ` ' 1

Residual standard error: 9.71 on 874 degrees of freedom

Multiple R-squared: 0.4137, Adjusted R-squared: 0.3949

F-statistic: 22.02 on 28 and 874 DF, p-value: < 2.2e-16

Размещено на Allbest.ru

...