Personal Statement Example (With Breakdown)

In my second year as an undergraduate, I ran a study on whether bilingual children performed differently from monolingual children on a visual search task in which the target was embedded in a cluttered background. The hypothesis, borrowed from a 2012 Bialystok paper, was that bilinguals would show a small advantage because of enhanced executive control. My results were unambiguous: there was no difference at all. The effect size was 0.03, and the confidence interval ranged from mildly negative to mildly positive. I was disappointed for about a week and then fascinated, because when I went back to the original literature I realized that the effect I was trying to replicate had been under serious methodological criticism for at least five years and I had simply not read the critical literature carefully enough. I had run a textbook study and ended up in the middle of a replication crisis. That experience shaped how I approach research. It is the reason I want to work on how cognitive claims — especially ones that travel fast through pop-science writing — are constructed, tested, and either validated or quietly dropped. My senior thesis, under Dr. Amelia Vance, extended this question into working memory training. I built a replication of the "dual n-back" study whose claims have driven a commercial training industry, and I found essentially the same pattern: the effect the original paper reported did not appear under preregistered conditions with appropriate controls. The paper is currently under review at a journal. I am applying to the cognitive science PhD program at Rutgers specifically because of Dr. Elena Ruiz's research on how methodological choices cascade through a literature and produce effects that look robust but do not replicate. Her 2023 paper on the publication trajectory of the bilingual advantage claim is the clearest articulation I have read of the problem I stumbled into as a sophomore, and I want to work on the formal modeling side of her project — specifically the question of how prior-probability assumptions in Bayesian replication analyses interact with publication bias. Dr. Ruiz's lab uses simulated literatures as a test bed for these questions, which is exactly the methodological stance I want to learn. I am also drawn to the program's required second-year coursework on computational modeling, which would give me the formal tools I am currently missing. My undergraduate training has been strong on experimental design and statistics but light on computational modeling, and the Rutgers program is unusual in making that sequence required rather than optional. I have been self-studying in preparation: I am about two-thirds through Lau & Krause's Probabilistic Machine Learning textbook, and I have implemented the Metropolis-Hastings sampler from chapter 11 in Python as a way to force myself through the math. My goal after the program is to work on replication and meta-science methodology either in an academic lab or at an organization like the Center for Open Science. The specific reason I am applying to Rutgers is that the combination of Dr. Ruiz's research direction and the program's computational modeling requirement gives me the clearest path into that career.