Explore the vast range of titles available.

Hong Kong’s franchised buses contribute significantly to its high pollution levels, which in turn result in excess mortality and hospitalizations. I show that replacing all of Hong Kong’s pre-Euro IV buses with cleaner Euro V buses would save 1260 statistical lives, among other benefits. The expected net benefit of such a project is HK$26.4 billion under a discount rate of 3.5%. This result is robust to a discount rate of 10% and a mortality decrease of half of what is expected. This is one of the first studies to estimate the public health impact of air pollution from franchised buses in Hong Kong. The city can therefore consider conducting accelerated replacement of its franchised bus fleet. Given the strongly positive result, other Asian cities with ageing bus fleets could also conduct similar benefit-cost analyses.

Leaderboards are a workhorse within the gamification literature. While the effect of a leaderboard has been well studied, there is much less evidence how one's peer group affects the treatment effect of a leaderboard. Through a pre-registered field experiment involving more than 1000 users on an online movie recommender system, we expose users to leaderboards, but different sets of users are exposed to different peer groups. Contrary to what a standard behavioral model would predict, we find that a user's contribution increases when their peer's scores are more dispersed. We also find that decreasing average peer contributions motivates a user to contribute more. Moreover, these effects are themselves mediated by group size. This suggests that existing theories of motivation and demotivation with regards to leaderboards may need revision, and also illustrates the potential of using personalized leaderboards to increase contributions.

Behavioral economics has grown immensely since Richard Thaler founded the field. In this thesis, I explore three different areas of behavioral economics. The first chapter explores how mood affects contributions to Wikipedia, an online public good. The second chapter explores whether self-restraint can be a driver of crime. The third chapter explores whether rational inattention in valuations can be a source of risk aversion.

We design and implement an adaptive experiment (a ``contextual bandit'') to learn a targeted treatment assignment policy, where the goal is to use a participant's survey responses to determine which charity to expose them to in a donation solicitation. The design balances two competing objectives: optimizing the outcomes for the subjects in the experiment (``cumulative regret minimization'') and gathering data that will be most useful for policy learning, that is, for learning an assignment rule that will maximize welfare if used after the experiment (``simple regret minimization''). We evaluate alternative experimental designs by collecting pilot data and then conducting a simulation study. Next, we implement our selected algorithm. Finally, we perform a second simulation study anchored to the collected data that evaluates the benefits of the algorithm we chose. Our first result is that the value of a learned policy in this setting is higher when data is collected via a uniform randomization rather than collected adaptively using standard cumulative regret minimization or policy learning algorithms. We propose a simple heuristic for adaptive experimentation that improves upon uniform randomization from the perspective of policy learning at the expense of increasing cumulative regret relative to alternative bandit algorithms. The heuristic modifies an existing contextual bandit algorithm by (i) imposing a lower bound on assignment probabilities that decay slowly so that no arm is discarded too quickly, and (ii) after adaptively collecting data, restricting policy learning to select from arms where sufficient data has been gathered.

We design and implement an adaptive experiment (a ``contextual bandit'') to learn a targeted treatment assignment policy, where the goal is to use a participant's survey responses to determine which charity to expose them to in a donation solicitation. The design balances two competing objectives: optimizing the outcomes for the subjects in the experiment (``cumulative regret minimization'') and gathering data that will be most useful for policy learning, that is, for learning an assignment rule that will maximize welfare if used after the experiment (``simple regret minimization''). We evaluate alternative experimental designs by collecting pilot data and then conducting a simulation study. Next, we implement our selected algorithm. Finally, we perform a second simulation study anchored to the collected data that evaluates the benefits of the algorithm we chose. Our first result is that the value of a learned policy in this setting is higher when data is collected via a uniform randomization rather than collected adaptively using standard cumulative regret minimization or policy learning algorithms. We propose a simple heuristic for adaptive experimentation that improves upon uniform randomization from the perspective of policy learning at the expense of increasing cumulative regret relative to alternative bandit algorithms. The heuristic modifies an existing contextual bandit algorithm by (i) imposing a lower bound on assignment probabilities that decay slowly so that no arm is discarded too quickly, and (ii) after adaptively collecting data, restricting policy learning to select from arms where sufficient data has been gathered.

We conduct a study of hiring bias on a simulation platform where we ask Amazon MTurk participants to make hiring decisions for a mathematically intensive task. Our findings suggest hiring biases against Black workers and less attractive workers and preferences towards Asian workers female workers and more attractive workers. We also show that certain UI designs including provision of candidates information at the individual level and reducing the number of choices can significantly reduce discrimination. However provision of candidates information at the subgroup level can increase discrimination. The results have practical implications for designing better online freelance marketplaces.

Neutrophils are dynamic with their phenotype and function shaped by the microenvironment, such as the N1 antitumor and N2 pro-tumor states within the tumor microenvironment (TME), but its regulation remains undefined. Here we examine TGF-β1/Smad3 signaling in tumor-associated neutrophils (TANs) in non-small cell lung carcinoma (NSCLC) patients. Smad3 activation in N2 TANs is negatively correlate with the N1 population and patient survival. In experimental lung carcinoma, TANs switch from a predominant N2 state in wild-type mice to an N1 state in Smad3-KO mice which associate with enhanced neutrophil infiltration and tumor regression. Neutrophil depletion abrogates the N1 anticancer phenotype in Smad3-KO mice, while adoptive transfer of Smad3-KO neutrophils reproduces this protective effect in wild-type mice. Single-cell analysis uncovers a TAN subset showing a mature N1 phenotype in Smad3-KO TME, whereas wild-type TANs mainly retain an immature N2 state due to Smad3. Mechanistically, TME-induced Smad3 target genes related to cell fate determination to preserve the N2 state of TAN. Importantly, genetic deletion and pharmaceutical inhibition of Smad3 enhance the anticancer capacity of neutrophils against NSCLC via promoting their N1 maturation. Thus, our work suggests that Smad3 signaling in neutrophils may represent a therapeutic target for cancer immunotherapy. TGF-β stimulated tumor-associated neutrophils (TANs) can exert pro-tumoral functions. Here the authors show that Smad3 activation in TANs is associated with an N2-like polarization state and poor outcome in patients with non-small cell lung carcinoma and that Smad3 targeting reprograms TANs to an antitumor state suppressing tumor growth in preclinical lung cancer models.

Transforming growth factor-β (TGF-β) was originally identified as an anti-tumour cytokine. However, there is increasing evidence that it has important roles in the tumour microenvironment (TME) in facilitating cancer progression. TGF-β actively shapes the TME via modulating the host immunity. These actions are highly cell-type specific and complicated, involving both canonical and non-canonical pathways. In this review, we systemically update how TGF-β signalling acts as a checkpoint regulator for cancer immunomodulation. A better appreciation of the underlying pathogenic mechanisms at the molecular level can lead to the discovery of novel and more effective therapeutic strategies for cancer.

Amomum villosum Lour., produced from Yangchun, Guangdong Province, China, is a Daodi medicinal material of Amomi Fructus in traditional Chinese medicine. This herb germplasm should be accurately identified and collected to ensure its quality and safety in medication. In the present study, single nucleotide polymorphism typing method was evaluated on the basis of DNA barcoding markers to identify the germplasm of Amomi Fructus. Genomic DNA was extracted from the leaves of 29 landraces representing three Amomum species (A. villosum Lour., A. xanthioides Wall. ex Baker and A. longiligulare T. L. Wu) by using the CTAB method. Six barcoding markers (ITS, ITS2, LSU D1-D3, matK, rbcL and trnH-psbA) were PCR amplified and sequenced; SNP typing and phylogenetic analysis were performed to differentiate the landraces. Results showed that high-quality bidirectional sequences were acquired for five candidate regions (ITS, ITS2, LSU D1-D3, matK, and rbcL) except trnH-psbA. Three ribosomal regions, namely, ITS, ITS2, and LSU D1-D3, contained more SNP genotypes (STs) than the plastid genes rbcL and matK. In the 29 specimens, 19 STs were detected from the combination of four regions (ITS, LSU D1-D3, rbcL, and matK). Phylogenetic analysis results further revealed two clades. Minimum-spanning tree demonstrated the existence of two main groups: group I was consisting of 9 STs (ST1-8 and ST11) of A. villosum Lour., and group II was composed of 3 STs (ST16-18) of A. longiligulare T.L. Wu. Our results suggested that ITS and LSU D1-D3 should be incorporated with the core barcodes rbcL and matK. The four combined regions could be used as a multiregional DNA barcode to precisely differentiate the Amomi Fructus landraces in different producing areas.

Conventionally, a soil nail is a piece of steel reinforcement installed inside a hole drilled into the slope and filled with cement paste (CP) grout. Chloride penetration is a major deterioration mechanism of conventional soil nails as the CP grout is easy to crack with an uncontrollable crack opening when the soil nail is subject to loading or ground movements. Engineered Cementitious Composites (ECC) are a class of fiber-reinforced material exhibiting excellent crack control even when loaded to several percent of strain, and therefore, ECCs have great potential to replace traditional CP grout in soil nails for achieving a long service life. In this study, the chloride ion transport characteristics and electrically accelerated corrosion process of steel rebar in ECC and CP grouts are systematically studied. The rapid chloride ion penetration test results showed a reduction of 76% and 58% passing charges in ECC with 0.15% and 0.3% pre-loading strain, respectively, as compared to that in un-cracked CP. Furthermore, the accelerated corrosion experimental data showed that ECC under pre-loading strain still exhibited a coefficient of chloride ion diffusion that is 20–50% lower than CP grout due to the ability to control crack width. Service life calculations based on experimentally measured parameters showed that the predicted corrosion rate and corrosion depth of soil nails in ECC grout were much lower than those in CP grout. The findings can facilitate the design of soil nails with excellent durability and long service life.