Sam Winn: Class of 2023

About Me: I am a senior at Princeton High School, following the three-year research track. My main research interests lie in big data exploration, but I have dabbled with applicable lab work including biological experiments and particle physics. I am currently working on constructing greenhouses for a sustainability project that will be entered as Princeton High School’s entry for the Samsung Solve for Tomorrow challenge! My favorite subjects are math and economics, and I plan on pursuing a dual-degree in economics and data science in college. My other hobbies include exploring the outdoors by hiking, canoeing and camping!

You can reach me at:, or (609)-608-8854

Project I: The Effect of Geographical Region on Race Inequality for Hiring Situations

(link to specific webpage:

Abstract: Our proposed stratifying inequality project will attempt to highlight, identify and understand existing biases and discrimination in the context of employee evaluations. We will attempt to identify gaps in ratings across four races (White, Black, Latinx, Asian) through a survey using MTurk (crowdsourcing marketplace)​. A total of 1,600 responses were collected, yielding ratings for 8,000 fictitious applicants. It was hypothesized that survey takers located in the Midwest and South would rate Black and Latinx applicants lower than survey takers located in the Northeast and West, due to the impact of predominant political parties in the regions (Quillian et. al, 2020). It was also hypothesized that survey takers located in the Midwest and South would rate women relatively lower than men in quantitative, male-dominated fields due to pre-established gender bias in the political parties of the region (Heilman, 2001). In our preliminary data analysis, the Midwest region rated Black candidates an average of 6% lower than the national average, and the Midwest rated Latinx candidates an average of 10% lower than the national average. Meanwhile, the Northeast rated Black and White applicants of equivalent credentials exactly the same. The Midwest rated White applicants the highest out of the four racial groups. Negative employee evaluations can significantly impact one’s ability to receive promotions and bonuses, preventing employee’s ascension on the socio-economic ladder. A deeper understanding of the biases that are laced in employee evaluations, along with the driving motives behind them, can help organizations reduce (and eventually eliminate) workplace discrimination.

Figure I: Sample graph showing disparity between Black, Chinese and Latino applicants when reviewed by participants in the Midwest region. Preliminary findings show statistically significant (p < 0.05) results that show these groups were discriminated against at a higher rate than white applicants.

Project II: Investigating Cognitive Mechanisms Behind Betta Fish Memory Capabilities, Dependent on Temperature

Abstract: Our project aims to determine the training and cognitive capabilities of two related fish species: Betta imbellis and Betta splendens. We will also attempt to identify the driving mechanisms behind risk/reward decision making in bettas by introducing both a food-based training regimen and the negative consequence of capture in a net. Over the course of several weeks, fish will be trained using food motivation techniques to follow a stick, culminating in the task of jumping through a hoop. 

The 6 phases of the experiment are the Looking Phase, the Following Stage, the Surface Stage, the Hoop Phase, the Hoop Jumping Phase, and finally the Potential Capture Phase. Fish will be trained with 10 reps per training session, and will advance to the next training phase after achieving a perfect rating for 3 consecutive sessions. The Potential Capture Phase seeks to test if the food motivation and previous training will be interrupted by the negative association of trauma/stress. In quantitative terms, the fish should achieve success on 9/10 reps after one week in each phase of training, and complete a minimum of 7/10 reps to demonstrate recall after a week-long gap in training. However, after being captured, the betta fish will only perform the hoop-leaping task on 2/10 reps, as the memory of stress will deter the fish from continued leaps. 

At this point, daily averages of the 9 fish scores have been taken, with a significantly increasing trend in the performance over a 10-day period. Furthermore, during all of the first trials, temperatures of the aquaria were recorded and a positive correlation between training performance and temperature were noted. The trends suggest that a decrease in temperature leads to an immediate decrease in the score that the fish achieved, and can conclude that warmer temperatures enhance the training performance of both Betta imbellis and Betta splendens.

Annotation: This graph depicts the aggregated training performance of all fish. It is clear that fish are able to learn, and that they have the ability to increase their knowledge over time. Separately, there is an increase in training performance around day 6 (the same day the temperature started increasing) Training performance flattens out as soon as the temperature stops increasing. 

Project III: Observation of Electron Drift in an Electrostatic Particle Accelerator and a CST

Researchers: Jonathan LaPoint, Niklas Austermann, Sam Winn

Our experiment aims to model the electron drift patterns through use of a physical e/m apparatus as well as a CST simulation of the same accelerator. The e/m apparatus, consisting of two coils opposite the tube each with an electrical current as well as two plates, one positive and one negative, propels electrons from the electron gun to the electromagnetic field present in the tube. Ionization with the argon gas inside of the tube produces a dim yet visible blue effect that is essential in tracing the path of the electron. Nevertheless, the experiment aims to further explore how alterations made to the electromagnetic field, caused by changes in the orientation of the surrounding coils, could change the electron’s path and drift.

Methodology (for the first part of the experiment)

The voltage of the particle accelerator’s power supply was set to 6.0 V with the current set to 1.5 A (DC is set to 250 volts). The base of the particle accelerator was modified using 3D printed pieces (design pictured below) in order to maximize the potential tilt of the Helmholtz coils to ± 15.0° in either direction. The 3D-printed turning mechanism, used to fix the coils to rotating bases secured to the secondary wooden base via a swivel, was attached to each coil. 

Then, 25% of the top plate was cut off and attached separately to the main top plate. The main top plate was reinserted into the apparatus with a tilt of approximately 45°, and the cut off section of the top plate was reinserted horizontally, centered and parallel to the tube. The apparatus was then turned on, and the BxB electron drift was observed.

With the CST simulation, we will attempt to model the tilt of the coils, and the separation of the top plate. We will observe whether our observed field results align with the results predicted by the simulation.

Figure I: The setup of the e/m apparatus described above, with the separated top plate and tilted coils. The top plate is fixed to the coils with duct tape but this is subject to change. At the moment, 3-D printed parts are the most feasible solution to securely fix the top plates.

Figure II: Results of the BxB electron drift from the setup depicted above. This image 

Figure III: We created a PVC frame to drape a black curtain over our apparatus, in order to obtain better photos. 

Figure IV: This was our photography setup prior to the addition of a PVC pipe framework. Sam is under the cloth, using a slow shutter app to get better lighted photos of the electron beam.