# Questions for BMSS discussion sessions

## Week 4

## Questions

* DellaPosta et al. (2015) mention Watts et al.'s (2006) "Music lab" study. In that study, Watts et al. showed how inferior music choices could become popular in the context of social influence. DellaPosta et al. (2015) interpret this as an instance where network autocorrelation can, over time, make "bad" songs become more popular than "good" ones. Another interpretation, however, is given in a paper by van de Rijt (2019), who states that Watts et al. stopped the studies too early — before the "good" songs could become more popular than the "bad" ones. He says that network autocorrelation is over time outpaced by self-correcting dynamics that make better choices also more popular. What do you think is the relationship between network autocorrelation and self-correcting dynamics in other situations (e.g., political polarization)? Are they complementary, and does self-correction always overcome network autocorrelation?

* Why are traditional random sampling methods problematic to account for network properties?

* DellaPosta et al. (2015) argue that network autocorrelation can explain network structure as a combination of observable, manifest variables (e.g., gender or ethnicity) and interactions between these variables and people's preferences. Can you use the mechanism of network autocorrelation to explain the findings of Leszczensky and Pink (2019)?

* DiMaggio (1997) speaks about culture and identities as a toolkit which people can use in order to navigate everyday life depending on the situation.
* Try to rephrase the basic idea of the Levina and Arriaga (2019) paper in terms of cultural toolkit: What is the "culture" there, and how is it being used by individuals in their model?
* Draw from Valdez (2015) and Bonomi (2019) to explain how factors external to individuals can trigger latent cultural traits to produce frame switching.

* Using Bourdieu’s concepts of field and capital(s) evoked in Levina, Arriaga (2014), explain the opposition between avant-garde and popular artists in cultural and artistic fields (literature, painting, fashion, cinema…).

### Maël

* Actions are guided by attitudes and beliefs. They are necessary but not sufficient to explain (voting) behavior. Draw from Valdez (2015) and Bonomi (2019) to explain what other factors determine voting behavior (public problem, framing, demographic changes…).
* Using Bourdieu’s concepts of field and capital evoked in Levina, Arriaga (2014), explain the opposition between avant-garde and popular artists in cultural and artistic fields (literature, painting, fashion, cinema…).

### Hendrik

* DellaPosta et al. (2015) mention Watts et al.'s (2006) "Music lab" study. In that study, Watts et al. showed how inferior music choices could become popular in the context of social influence. DellaPosta et al. (2015) interpret this as an instance where network autocorrelation can, over time, make "bad" songs become more popular than "good" ones. Another interpretation, however, is given in a paper by van de Rijt (2019), who states that Watts et al. stopped the studies too early — before the "good" songs could become more popular than the "bad" ones. He says that network autocorrelation is over time outpaced by such self-correcting dynamics that make better choices also more popular. What do you think is the relationship between these two mechanisms in other situations (e.g., political polarization)? Are they complementary, and does self-correction always overcome network autocorrelation?
* Bonomi et al. (2021) make a very strong assumption of their model: "We assume that there are only two groups with which a voter can identify: her economic group or her cultural group, but not both." (p. 2387). Given what you have read in the other texts, do you think this assumption can be defended in this case?
* Leszczensky and Pink (2019) find that German-Turkish people who do not strongly identify with German or Turkish culture (international in their words) tend to befriend other people who also do not strongly identify with any single culture. However, they also mention that these people stay away from people who strongly identify with an *international* identity. Why is that so? Wouldn't it make more sense to befriend people who value interculturality stronger?

### Alex

- Do you buy DellaPosta, Shi and Macy’s (2015) argumentation that the “self-reinforcing dynamics of homophily and influence dramatically amplify even very small elective affinities between lifestyle and ideology” leading to a world where aesthetic preferences become strongly connected to ideological or political profiles? Why do you think it is convincing, why not? Can you come up with an alternative social mechanism that might potentially explain the observed pattern?

- Many sociologists view culture as restricting, limiting people’s imaginations and aspirations to existing arrangements. Some research, however, indicates that people also act as if they use cultural elements as strategic means (DiMaggio, 1997). Can you think of social mechanisms that enable the one or the other? Which role do schemata, context, and group-memberships play?

- Levina and Arriaga (2014) outline a theory of status production on online user-generated content platforms. => Mechanisms "how social media design impacts interaction and social structure" ? (not done yet... I am lacking an idea how to put this)

## Week 3

## Questions

* What are the differences between conventions and heuristics? How do they relate to each other?

* What role plays social influence in the emergence of norms?

* Tversky and Kahneman became famous for introducing the idea of bounded rationality, in which they integrated psychological insights into economics and rational choice theory. Their main ideas are based on the assumptions that (1) "most judgements and choices are made intuitively" and (2) "rules that govern intuition are generally similar to the rules of perception" (Kahnemann, 2003, p.1450). Hence, they proposed an architecture of cognition consisting of three layers: Perception, Intuition (System 1), and Reasoning (System 2) (see Figure 1). Operations of system 1 are described as "fast, automatic, effortless, associative, and often emotionally charged; they are also gov- erned by habit, and are therefore difficult to control or modify", while operations of system 2 are "slower, serial, effortful, and deliberately controlled; they are also relatively flexible and potentially rule-governed" (Kahnemann, 2003, p.1451). Take the Healy and Lenz (2014) paper and try to identify the system(s) responsible for the behaviour they observe. Do you encounter limitations of Kahneman's theory? Are there elements that cannot be captured or are ambiguous?

* Oftentimes, different heuristics can be used to describe similar phenomena. Consider two examples from the course's readings. First, think about how Miyoshi et al (2021) analyze what drives people to become vaccinated versus how Centola et al (2005) attempt to explain how unpopular norms get enforced. As the second example, think about the "bunching" that Allen et al. (2017) were describing in last week's readings, vis-a-vis Frey and Goldstone's (2018) "flocking". How different are these examples from each other? Can you come up with a "meta heuristic" that generalizes the first and the second example? If not, why?

* Throughout the readings, one can clearly see that there are two general approaches to measuring heuristics: either by analyzing real-world-data, or by running a simulation. What are the benefits of each of the two approaches when it comes to explaining macro-outcomes from micro-level behavior? In order to think structurally about this, try to come up with a setup for a simulation that reproduces the Marathon runner example from last week (Allen et al. 2017). Conversely, try to come up with how one could measure the heuristics that Centola et al. in this week with real-world-data. How easy or difficult is this translation between simulation and data? Why?

### Hendrik

* According to Convention Theory, the "market" works as it does mainly because it is "embedded" — embedded in social norms and structures. However, sometimes conventions and market mechanisms can become decoupled, for example when basic necessities are no longer available or unreasonably expensive. Think of examples for heuristics, or conventions, that people could reasonably have with regard to supply of basic goods, which were crushed by the COVID-Pandemic. How have these conventions or heuristics been altered since? Are there now new conventions? (question about the Biggart & Beamish (2003) paper)

heuristics, conventions, norms

### Alex

- measurement strategies for heuristics (Hendrik & Mael have already covered this)

### Maël

- In their paper, Biggart and Beamish argue that conventions play a role in maintaining order in the economy. Provide examples of both standard and habitual conventions and explain why they participate in maintaining economic order.
- Relate that concept with that of social order and inequalities.

- In Centola et al. 2005, authors argue that an embedded network reduces the spread of the unpopular norm. Reflect on the role of network structures in political polarization.
- Several of this week’s readings include simulation methods. Explain why simulation is a good tool to investigate social influence mechanisms.
- Several of this week’s readings include experimental settings. Explain why experiments are a good tool to investigate behavioral mechanisms and highlight limitations of experiments.

## Week 2

## Questions

* RCT serves as one of many starting points to study micro-macro level links. Friedman and Hechter (1988) state that opportunity costs, institutional constraints, and preferences are the main source of variation in rational choice explanations among different individuals. Do you think that this covers all sources of variation? Which dimension could be missing? Take a look at Friedman and Hechter (1988), Figure 1 in order to structure your thoughts.

* Think about the relation of RCT with Analytical Sociology and Middle-Range Theories. Which circumstances make one theory more applicable to study social phenomena than the other?

* What is a model? What is its role in social science?

* Some of the articles for this part of the seminar implicitly or explicitly refer to the “free-rider” effect. Also, some refer to the “self-fulfilling prophecy”. At first glance, they seem contradictory: the free-rider effect states that, as the utility expectations for more and more people go up, actual participation in realizing the utility goes down, while the self-fulfilling prophecy only works when more and more people actually participate in realizing some public benefit. Do these two effects cancel each other out (i.e. the larger effect prevails)? Do they mutually reinforce themselves (i.e. as free-riding goes up, so does participation by others)? Can we even observe both effects at the same time?

* Schwartz et al. in their article find that maximizers generally are less happy than satisficers as it is difficult to find the optimal choice. However, what happens when there are either many or only a few choices? Think about two examples: deciding whether to stay in school (Breen & Goldthorpe) versus choosing a new favorite song. In the first example, the choice should be easy even for maximizers, while in the second example, it seems impossible to come up even with a “good enough” choice. Think about what mechanisms might help (or put at a disadvantage) both maximizers and satisficers in those two scenarios.

### Maël

* RCT often make assumptions about individual behavior. According to you, are they realistic? Justify with examples from the articles.
* Is it a better to maximise or to satisfice?
* What is a model? What is its role in social science?
* Can RCT be associated with middle-range theories? Take examples from the articles.
* Is finishing a marathon before round time a rational action? Is
* Field and rational action
* Emotion and rational action? cf. [weber](https://www.yourarticlelibrary.com/sociology/types-of-social-action-according-to-max-weber/43755).

### Hendrik

- Schwartz et al. in their article find that maximizers generally are less happy than satisficers as it is difficult to find the optimal choice, inducing negative emotional responses when faced with a decision. However, what happens when choices are more or less restrained? Think about two examples: deciding whether to stay in school (Breen & Goldthorpe) versus choosing a new favorite song. In the first example, the choice should be easy even for maximizers, while in the second example, it seems impossible to come up with even a “good enough” choice as art is generally thought to be subjective. Think about what mechanisms might help (or put at a disadvantage) both maximizers and satisficers in those two scenarios.
- Some of the articles for this part of the seminar implicitly or explicitly refer to the “free-rider” effect. Also, some refer to the “self-fulfilling prophecy”. On a first glance, they seem contradictory: the free-rider effect states that, as the utility expectations for more and more people go up, actual participation in realizing the utility goes down, while the self-fulfilling prophecy only works when more and more people actually participate in realizing some public benefit. Do these two effects cancel each other out (i.e. the larger effect prevails)? Do they mutually reinforce themselves (i.e. as free-riding goes up, so does participation by others)? Can we even observe both effects at the same time?
- Many companies and countries are currently trying out the so-called 4-day-week which promises employees only four days of work per week, while the pay remains the same. A five day week implies that the most utility of the work of an employee will be realized only after 35-40 hours. If employers now switch to 4-day-weeks (~28 hours a week), this utility theoretically should go down, but in fact the same utility value remains, according to the employers who have already trialed such a 4-day-week. Try to explain why that is by applying Allen et al.’s notion of “bunching” to this situation.

> [likewise: or "public goods game" <= Many articles refer implicitly, or explicity to this one, too.]
the “free-rider” effect
### Alex

* RCT serves as one of many starting points to study micro-macro level links. Friedman and Hechter (1988) state that opportunity costs, institutional constraints, and preferences are the main source of variation in rational choice explanations among different individuals. Do you think that this covers all sources of variation? Which dimension could be missing? Take a look at Friedman and Hechter (1988), Figure 1, so strutcure your thoughts.

* Compare RCT with Analytical Sociology. Where do you think both approaches go hand in hand? Where do you see differences? Which circumstances make one theory more applicable to study social phenomena than the other?

* Although classical RCT utilises stong assumptions, that not necessarily hold in reality, do you see value in this approach? What would this value be?

+ Which "mechanisms for rationality" do people utilise? How do they deal with real-life complexity?

## Week 1

## Questions

- What is a middle-range theory and how it is linked to grand theory (look it up if necessary)?

- Why do you think it is important to look at mechanisms?

- What are the problems of the covering-law model and how can they be solved with a mechanism-based approach?

- Which purpose does the Coleman diagram serve?

- In Ylikoski (2016), the author updates the Coleman diagram with a feedback loop. What is the rationale behind this fifth arrow? Find a social phenomenom that includes this feedback loop and explain all arrows and nodes.

- What is the relationship between methodological individualism and structural individualism?

---

### Maël

- According to Hedström, what is the difference between mechanisms and laws? why are mechanism-based explanations preferable in social sciences?
- Similarly, what is the difference between the “covering-law” and mechanism-based approaches? why are mechanisms better tools to explain social phenomena?
- To what fallacious reasonings can a “covering-law” type of explanation can lead to?
- Is the mechanistic approach assuming methodological individualism? Why? (tricky question).

### Hendrik

- [GREAT] Hedström and Swedberg mention that, while social mechanisms cannot take the form of deterministic laws, they can take the form of “probabilistic laws”. However, one paragraph later, they explicate an important constraint on that: these laws should not take the form of mere “statistical associations”. So what would a mechanism look like that is a “probablistic law” while not mirroring a mere statistical association? Is this even possible?
- According to Hedström and Ylikoski, there are two kinds of explananda as part of mechanisms: one regarding empirical facts, and one regarding stylized facts. What is the difference between the two? Can you give an example?
- [GREAT] There are four principles at the core of a mechanism-based approach. However, three of them seem to be at odds at first: precision versus abstraction and reduction. Upon closer inspection, they complement each other, but how is that possible?
- At several points, the authors state the centrality of simulations for mechanism-based explanations. However, there are many scholars out there who develop mechanism-based explanations without the help of simulations. What is the central property that an analytical strategy needs to be mechanism-based, and why?
- Ylikoski mentions that the Coleman tub should be seen more as a tool for thought rather than a rigid structure, but the authors do agree that a mechanism-based explanation needs to consist of all three arrows. Can you give an example for a (potential) mechanism that follows this structure and think about how one would go about checking each of the three arrows?

### Alexandra

Ylikoski text (2016)

- What is the role of sociology/social science for social inquiry?
- Which puspose does Coleman's diagram serve?
- What are common missinterpretations of the Coleman boat/diagram?
- How could the Coleman boat be extended in meaningful ways?
- Which implications does it have that Coleman accepts cooperate and collective agency?

Hedström, Swedberg (1998) text

- Write down your definition of a mechanism.
- Why do you think it is important to look at mechanisms? Which role does “mechanism-based” thinking play for social science inquiry?
- What are situations where mere descriptions can be useful for social science inqiry?
- Which other approaches might be available to us? How can they work together?

### Sven

https://liuonline-my.sharepoint.com/:w:/g/personal/svest481_student_liu_se/EbU6QFBN7NlGplxGgacIMXIBhC1rKQYL7lyq3XEjZL9iiQ?e=9NkAkN