2018-2019 Fall Colloquia

ABSTRACT: A vast body of literature documents important differences in the politically salient attitudes and behaviors of those who do and those who do not attend college. However, while this body of research has employed rigorous causal inference methods, the findings about the political effects of college have been highly contradictory: while some find that college is a source of strong political (re)socialization, others find that the differences between college goers and non-goers simply reflect the selection of certain types of individuals into college. A casualty of such hyper focus on (dis)proving causality is that research on the process through which college contributes to such outcomes has been limited. A better understanding of the mechanisms through which college could affect people at all can help elucidate why the so-called “college effect” is so elusive. The point of departure of this paper is that the effect of college on sociopolitical attitudes is contingent on social contexts, such as interactions among students, which shape motivations for attitude change. I use an in-depth case study of a cohort of students at one private university in the US that observed students starting the summer before college to the end of their graduating year, and use students’ affiliation networks to understand the sociopolitical climate they are exposed to within the larger organizational setting each year. Using identity theory as a lens to understand how groups influence individuals’ beliefs and perspectives, I show that changes in individuals’ political self-identification is predicted by network processes rather than by any macrocultural influence of the college as whole. This research suggests that college is not a “uniform dose” treatment, and that research on college effects can benefit by taking social networks of colleges into account.

ABSTRACT: A “resignation industry” is developing in tandem with—and overlapping with—the growth of the digital interactive media system. The resignation industry carries out pervasive and purposeful corporate activities that encourage people to give up thinking they can change data collection by businesses. These undertakings have the potential of corroding political and cultural democracy. We need a sociology of digital resignation to understand the industry. Research in this area is best carried out with a new understanding of the meaning and nature of “media research.”

ABSTRACT: The recent reports of the Commission on Evidenced-Based Policymaking and the National Academy of Science Panel on Improving Federal Statistics for Policy and Social Science Research Using Multiple Data Sources and State-of-the-Art Estimation Methods emphasize the need to make greater use of data from administrative and other processes. The promise of such data sources is great, and even more so if multiple data sources are linked in an effort to overcome the shortage of relevant information in each individual source. However, looking at countries in which administrative data have been accessible for longer, or the tech industry in which process data are used extensively for decision making, we see that process data are often insufficient to answer relevant questions or to ensure proper measurement. This creates a desire to augment process data and administrative data with surveys. This talk will focus on two practical aspects resulting from this situation: the enormous challenge in ensuring privacy, and the need to cross-train computer scientists, statisticians, and survey methodologists.

ABSTRACT: The accuracy of the results from any survey depends on minimizing nonresponse bias, which consists of systematic differences between people who respond to a survey and people who do not. In this talk, I present preliminary results of an ongoing analysis of non-response bias in the 2016 American National Election Study (ANES -- a nationwide survey of adult US citizens conducted using both face-to-face interviews and Internet questionnaires). First, I compare the accuracy of ANES estimates to known population characteristics. This reveals alarming differences. Next, I compare "easy to get" ANES respondents (such as those responding on the first invitation) to "hard to get" respondents who required multiple invitations or refusal conversion. These comparisons show the results of extraordinary field effort to interview respondents who were difficult to contact or reluctant to cooperate. They reveal the effects a lower response rate would have had on the accuracy of estimates, and they let us make an educated guess about the effect a hypothetical higher response rate would have had on accuracy. Third, I present the results of an exceptional resource in non-response bias analysis: a non-response followup (NRFU) study. The NRFU allows direct comparison of ANES respondents to (some) ANES non-respondents. Finally, I discuss the results in terms of their implications for the evaluation of data quality and for operational design of future data collections in both face-to-face and Internet modes, including effective recruitment, refusal conversion, and adaptive design to counteract non-response bias.

ABSTRACT: The increasing urgency to address climate change drives us to consider ways to motivate personal and public action beyond emphasizing that climate change is real. One promising approach has been to focus on the co-benefits for society of addressing climate change, such as positive outcomes for health, technological development, and the economy. But when is this focus likely to work? I will describe a framework we developed to understand these co-benefits. I identify the types of co-benefits that are more likely to motivate action, and how to focus on these co-benefits in communication to increase their motivational impact. In this latter work our findings challenge some of the basic assumptions about how to address temporal dimensions in communicating climate change.

ABSTRACT: In the past two decades, increasing levels of simplicity in political elite rhetoric have drawn both empirical interest and normative concern from political scientists and communication scholars. While the conventional wisdom holds that politicians simplify their public communication because “simplicity works,” the way citizens respond to such messages is unclear. I present the results of two experiments testing the effects of simple rhetoric on citizens' attitudes. Results point to a conditional effect. When a politician addresses a like-minded audience, simplicity sways public opinion. However, when addressing a polarized audience, simple rhetoric is ineffective.

ABSTRACT: In the past few decades, consumer researchers have studied consumer decision making using primarily relatively small choice sets that were described in terms of two or only a few attributes. These studies have provided a wide range of insights, such as regarding context, task, and preference elicitation effects on consumer choice. However, today’s consumer choice environment has dramatically changed, in large part due to the rise of the Internet and related developments. In particular, consumers today have easy access to numerous options, which they typically consider sequentially.

I will present two projects that focus on the manner in which consumers evaluate options in the current environment, the evolution of consideration sets (or “shortlists”), and the impact of the manner of evaluation on purchase likelihood. The first project is conducted in collaboration with Jen Park (a current doctoral student). Jen and I are proposing that the framing of decisions to not include an item in a consumer’s final choice set can affect her commitment to the constructed choice set. We test this idea in the online context, where the experience of adding items to the consumer’s cart impacts the way the consumer ultimately evaluates the items in her cart and makes (or does not make) a purchase. In a series of studies, we allow participants to browse through consumer products and add items to their shopping cart for reconsideration. When asking participants whether they would like to add the product to their cart, we vary the framing of decisions to pass on the product as either “reject this item” (which can be seen as a hard rejection) or (“look at other options” (a soft rejection) and examine its downstream psychological and behavioral consequences.

The second project is conducted with Wendy Liu (U.O California, San Diego). In this research, Wendy and I examine shortlisting (or consideration set construction) behavior when options are added to the shortlist sequentially. For example, rather than wait until seeing all available wines to simultaneously evaluate all options, the consumer keeps a running shortlist, perhaps for fear that one will not be able to remember the information about all wines.

We propose and demonstrate that people tend to have a metacognitive preference for a smaller shortlist, in order to make the final choice easier for themselves. Given this goal, the person self-regulates the size of the shortlist in making on-the-go shortlist decisions, resulting in shifts in preferences. It is as if they impose a cost on the addition of another option to the shortlist, which must be weighed against the benefit of adding the option in terms of expected gain in eventual choice outcome. Importantly, we propose that the salience of this cost is not constant; instead, it increases as the shortlist grows. That is, in the beginning of the sequence, when the shortlist is small, the goal of limiting the shortlist size is less salient, and thus there is little hesitation in adding an option to the shortlist (assuming the option is deemed attractive enough). However, when the shortlist becomes larger, the issue of having an eventual shortlist that is too large becomes more salient, and the threshold for adding options rises. Thus, as the shortlist grows, the current option under evaluation faces increasing resistance to be also added to the shortlist.

ABSTRACT: Communication is complex and dynamic processes involving both exogenous and endogenous influences. In this talk, I will describe a theoretical framework and research paradigm called Dynamic Motivational Activation (DMA), which uses real-time data (e.g., psychophysiological measures, longitudinal experience sampling data) in conjunction with formal dynamic models to understand how people attend to, respond to, and select media in an adaptive way, and how these processes interplay with and impact our physical and psychological wellbeing. The DMA approach helps tease apart the influences of the exogenous variables (e.g., message content and design variables) and the endogenous variables (e.g., feedback effects of audience’s physiological, cognitive, and behavioral systems), and allow the study of their dynamic interactions over time. Several applications will be discussed, including identifying patterns of real-time attentional and emotional responses to media messages, and the impact of media multitasking on our wellbeing.