Multisensory Statistical Learning: Modality General or Specific?

Statistical learning (henceforth SL) is the unconscious acquisition of structured information from environmental inputs. For example, we can extract word boundaries from the frequency of syllable transitions. SL is a multisensory phenomenon, having been demonstrated in multiple modalities and stimulus domains. However, it remains to be seen whether SL is supported by independent modality-specific mechanisms (i.e. a separate mechanism for vision, hearing, and touch) or a singular modality-general mechanism. In the present study, we test these models by manipulating the rate of presentation for auditory and visual SL. If the mechanism is modality-general, the rate of presentation should not affect learning and performance will be similar across modalities. If the mechanisms are modality-specific, the rate of presentation should affect learning: subjects should learn auditory sequences better at faster presentation rates and visual sequences at slower rates. Preliminary results suggest that auditory and visual sequences are learned most effectively at different rates of presentation, providing evidence of a ‘modality constraint’ on SL. This suggests that multisensory SL may be supported by modality-specific mechanisms, though further research and data collection are required

Parallel Planning and Verb Tense Errors

Speech errors are a common phenomenon, even in one’s native language. Thus, speech errors have been widely researched in psycholinguistics as a way of understanding the mental processes involved in language production. In this study we investigated the type of errors that occur when a speaker uses a tensed form of a verb when a non-tensed form should have been used. For example, in the sentence “She didn’t meant that,” the tensed form “meant” should be the non-tensed “mean”.
One cause for this type of error is interference from parallel sentence plans, where speakers may be planning multiple sentences in parallel, and errors may arise when elements from a competing sentence leak through. (Brehm et al, 2022).
An example of parallel planning is when a speaker begins activating both “No problem” and “One moment”. They might start responding “One [moment]”, but if the other response is still active, they might actually end up saying “One problem.”
In order to investigate this effect with verb tense errors, we conducted an online behavioral study of English speakers using audio data that participants recorded on their own computers. We manipulated factors such as past/present tense, ir/regularity and non/connectedness to nouns in the sentences to determine whether any of these factors made tense errors more likely. We are in the process of analyzing the data. The results will help us better understand the cognitive processes underlying verb tense production and sentence planning in general.

Comparative Analysis and Visualization of Campus Sexual Assault Data

Verb agreement avoidance
Why is it that humans use language the way we do? Humans communicate in numerous ways but one of the most significant components is language. We rely on the words we use and the sentences we compose to communicate but, we also avoid saying certain things, due to taboo or error avoidance. Previous experimental linguistic work has looked for agreement errors and ignored verb agreement avoidance. For example, “the label on the bottles are sticky” would be analyzed as an error, whereas, “the label on the bottles had been sticky” has no agreement at all, therefore, would be thrown out and assumed to have no impact on the data in an experiment.
It is known that humans avoid verbs, the real question is why do we avoid them? We examined data from agreement experiments that had already been collected, looking for patterns of verb agreement avoidance, and whether avoidance occurred in conditions where participants showed evidence of planning difficulty. We determined that difficulty planning does not correlate with verb agreement avoidance, but verb agreement avoidance is more common with certain conditions, such as when the verb comes after an adverb. We are still trying to understand why agreement avoidance happens, and what it means.

Defining the Conditions for an Effective Transition to a Decentralized Currency Standard in Developing Economies

In an increasingly digitized world, there is currently a significant shortage of literature on the benefits and drawbacks of a digital economy, particularly in regards to a national digital
currency standard. This research is designed to research the monetary policy measures designed to establish a Central Bank Digital Currency that fosters greater economic growth and inclusion within developing economies. Economic growth, in this instance, can be defined as increased rates of individual participation within an economy as well as a greater per-capita income for individuals within the economy. The selected method of inquiry for this research is a literature review of existing literature pertaining to individual topics surrounding how digital currency currently functions in developing economies, and how monetary policy could potentially affect a
digital currency standard. Findings from this project include: nationalized digital currencies would allow for more efficient transmission of remittances, Bitcoin differs from a CBDC from a
policy management perspective but could be fairly similar in regards to a technological/manufacturing perspective, and improved individual financial participation in
developing countries as a result of the use of a digital currency is yet to be empirically proven. While these findings are significant, they are largely a baseline of information to inform future
studies in the field. As a result of digital currencies being such a new development for all economies, and the limitations on that country’s economy or economic data as a result of its
political situation, statistically significant findings about the effects of national digital currencies are few and far between. Future research pertaining to the project aspires to provide an empirical correlation between the use of digital currency and improved individual financial participation, resolving the third finding of this study.

Using Global Value Chain analysis to study the current structure of the Global
Semiconductor Industry

Despite the swath of media coverage on the Semiconductor Industry and the shortages of
semiconductor components experienced over the past two years, conventional dialogue often
neglects to consider the heterogeneity within this sphere of industrial activity. To address these
shortcomings, this paper offers a Global Value Chain (GVC) framework by which to model the
current structure of the Semiconductor Industry. Section 1 provides a synopsis of semiconductor
components, the industry responsible for their production, and proposes a GVC framework by
which to evaluate both. Sections 2, 3, and 4 outline the input-output structure of the value chain,
the key players that operate within the industry, and the geography of the value chain,
respectively. Section 5 equips and applies this model to discuss three highly relevant topics: the
end-use demand for semiconductors, the economics of foundries, and the industry’s sales cycle.
Section 6 builds upon these applications as well as the GVC framework to examine the causes of
and implications behind the recent global semiconductor shortage. Section 7 offers our
concluding remarks on these matters as well as our key takeaways.

Women Leaders in the Baltic States: Untying the Double Bind

ABSTRACT: Across the post-Soviet region, but particularly in the Baltic states, women executives have gained power in greater numbers and at higher rates than many other regions in the world. This defies existing literature, as these states maintain conservative gender stereotypes while also facing a major security threat from Russia close to their borders. This thesis posits that the increase in women within Baltic legislatures across time creates a political pipeline, or a pool of qualified candidates that makes the election of women to executive power more likely. This is not the only factor, however, as the influence of NATO as a guarantor of Baltic security cannot be understated. Thus, this research finds evidence that NATO’s Article V has provided a security guarantee for the Baltic states, which has resulted in a consensus around security policy. This removes security as a major issue in elections, and lessens the burden of proving security competence from women leaders – an area which often aggravates the effect of stereotypes. If this security consensus is disrupted, however, women face an altered double-bind scenario in which they must balance the maintenance of a positive relationship with NATO/the U.S. while fulfilling their cultural role as women, tied deeply to national survival and independence. Estonia and Lithuania are selected as case studies, which culminate in an examination of current Estonian Prime Minister Kaja Kallas and Lithuanian Prime Minister Ingrida Simonyte.

Since college women are particularly vulnerable to sexual violence, it is essential to understand the knowledge and beliefs that students have surrounding Campus Sexual Assault (CSA). Results revealed that female students were more educated on sexual assault than male students. Additionally, fraternity men were more likely to reject rape myths than non-affiliated men. Our findings demonstrate the need for sexual assault education programs to go beyond consent knowledge, including an exploration of misogynistic attitudes and behaviors. Improvements to sexual assault education and more research regarding rape myths will help diminish rape-supportive attitudes, and subsequently decrease the prevalence of CSA.

Sexual Violence and Greek Life in the time of COVID-19

Greek life is a predictor of engagement in hookup and party culture, which are associated with campus sexual assault. COVID-19 mitigations may have diminished the impact of Greek Life on victimization by limiting exposure to Greek events. Our findings indicate sexual violence is still more frequent among Greek affiliated students, particularly women. This and other literature suggests that Greek male-dominated party infrastructure may have continued to promote community victimization in spite of pandemic public health efforts.

COVID-19 and Collaboration: How Maryland Expanded its EITC
The Earned Income Tax Credit (EITC) is one of the most important anti-poverty programs. Yet, eligibility for the federal EITC is dependent upon filing taxes with a Social Security Number. This means that people who file taxes with an Individual Tax Identification Number (ITIN), a group which largely consists of noncitizens, are ineligible for this benefit. However, several states have recently extended eligibility for state EITC benefits to ITIN filers. Drawing upon exploratory research in Maryland, which expanded EITC eligibility to ITIN filers in 2021 through passage of SB0218, I set out to understand the process, intentions, and effects of extension of state welfare programs to noncitizens. For this research, I conducted eight semi-structured interviews with policymakers and members of advocacy organizations that were involved in the expansion of the EITC. Based on this research, I argue that the context of the COVID-19 pandemic and collaboration between invested individuals and groups both heightened and highlighted the need for ITIN filers to be included in the Maryland EITC. The context of the pandemic created a necessary sense of urgency for politicians to get a bill passed. However, what ultimately caused the bill’s success was the strategic collaborations between policymakers, advocacy organizations, and different caucuses which took advantage of the pandemic context. The information learned here can help in future studies looking at the policy processes in other states and understanding if it is replicable in states that have not yet expanded EITC eligibility to ITIN filers.