InSci: Bridging the Gap in Scientific Literacy for High School Students
Client:
High School Students
Year:
2023 (3 months)
Skills Used:
Product Design (UX & UI)
Wireframing
Prototyping
User Research
Usability Test
Toolkit:
Figma
Miro
Canva
Github
“Poke” and InSci
Have you ever been excited about a topic but found yourself stuck trying to understand it?
That was me as a high school student, completely fascinated by developmental psychology but blocked by the complex jargon in scientific literature.
That's why I created cute "Poke." Poke isn’t just me; he represents every student eager to dive into the world of science but held back by confusing language and a lack of support. Poke is a symbol of young scientific enthusiasts whose passion needs to be protected and nurtured.
The heart of this project is to break down those barriers, simplifying the journey so that students like Poke can continue their scientific adventures without losing their spark. Poke’s story is about preserving that curiosity—making sure it thrives, not fades.
Background
Problem
Despite the essential nature of reading primary research literature for both scientists and students in science degree programs, many high school students struggle with comprehending scientific literature due to its complex language and intricate details. This challenge hinders their engagement with scientific research and prevents the development of crucial skills needed for scientific careers. The difficulty is exacerbated by the limited support available to help students navigate and interpret scientific texts effectively.
Rationale
Engaging with scientific literature is a critical skill that underpins successful scientific careers. Research shows that students at different educational stages, from undergraduates to faculty academics, exhibit varying levels of confidence and skill in reading scientific papers (Shanahan, Fisher, & Frey, 2022). The complex language, specialized vocabulary, and intricate details of scientific texts present significant barriers, especially for inexperienced readers such as high school students and early career scientists (Fang & Schleppegrell, 2010). These barriers prevent students from deeply engaging with scientific research, reducing their motivation and potential to pursue science-related fields (Wang & Adesope, 2016).
Project Goal
Create an AI-powered scientific research assistant that enhances high school students' comprehension and engagement with scientific literature.
The Process
01/Empathize
02/Define
03/Ideate
04/Prototype
05/Test
01/Empathize - Exploring the User’s Needs
Competitive Analysis Findings
The four top competitor apps in the educational technology space offer a range of valuable features similar to those of InSci, such as comprehensive educational content, interactive learning tools, and personalized study aids. However, none of these competitors provide a focused approach to simplifying complex scientific literature specifically for high school students, combined with interactive research assistance and real-time personalized guidance. This analysis reveals a significant gap in the market, aligning with our hypothesis that InSci's unique combination of AI-driven simplification of scientific texts, inspiring question generation, and tailored learning pathways addresses unmet needs in the current educational landscape.
User Interviews
To better understand the pain points and unmet needs of high school students who engage with scientific literature, I conducted in-depth interviews with 5 high school students who regularly engage with scientific literature as part of their school curriculum. The following is a summary of their experiences:
Challenges
All participants indicated they frequently struggle with understanding the dense, technical language used in scientific literature. Four out of five students find the jargon and complex grammar frustrating and disengaging. Despite this, they all want to improve their understanding of science for academic and career goals. One high achiever mentioned a lack of resources that simplify content as a major hurdle.
Research Process and Resources
Every participant undertakes some form of research to better understand scientific literature. However, their approaches vary. Some prefer to look up simplified summaries and explanations on educational websites like Khan Academy and YouTube. They follow educators and science communicators who break down complex topics into simpler terms. Others resort to reading multiple sources, including textbooks and articles, to piece together a coherent understanding. One participant, who often struggles with independently understanding these texts, relies heavily on group study sessions and discussions with peers and teachers. Despite these efforts, all participants expressed a need for more streamlined and reliable resources that can simplify scientific texts while maintaining accuracy.
Learning and Resource Preferences
Students prefer interactive content like quizzes and visual aids to reinforce learning. Step-by-step explanations and real-world applications are also valued. They appreciate platforms that offer a variety of content formats, integrating videos, articles, interactive quizzes, and peer-reviewed papers. Cost is important; free resources are preferred, but there’s a willingness to pay for high-quality tools. One student emphasized the need for comprehensive data and case studies in environmental science.
Overall Insights
Students are eager to enhance their scientific literacy but are hindered by complex language and lack of comprehensive resources. They need tools that simplify texts, offer interactive learning, and provide personalized guidance. Integration of multiple content formats and practical applications of theories are also important.
Ethical Matrix
Before designing InSci, I utilized an ethical matrix to systematically evaluate the impact of the tool on various stakeholders, ensuring that their interests and values were comprehensively addressed. The ethical matrix covered five key areas: privacy, equity, transparency, accountability, and educational integrity.
Key Takeaways: It underscored the importance of maintaining user privacy, ensuring equitable access to AI guidance, and providing transparent communication about AI functionalities and data usage. Accountability mechanisms were deemed essential to align AI recommendations with educational standards, while upholding educational integrity to support rather than undermine independent student research.
02/Define- Establishing the User’s Needs and Problems
Persona
Based on patterns in user behavior and common traits from the user interviews, I created two personas to help focus upcoming design efforts (While our primary target audience is high school students, teachers play a pivotal role in guiding students' learning experiences and often seek tools that can help simplify complex scientific concepts for their students, so I also created a persona of high school science teachers to help the tool effectively supports educators in enhancing student learning by understanding their needs, challenges, and behaviors).
Keeping in mind InSci’s business goal of targeting high school students & science teachers with an online solution that simplifies scientific literature, these personas represent the needs and behaviors of the users from that particular demographic.
Problem Statement
To understand and address the challenges faced by our target users, I conducted an exercise to generate "How Might We" statements. This process allowed me to identify key areas where I could focus our design efforts to improve the user experience with InSci. From approximately 20 "How Might We" statements, I decided to concentrate on the following three:
How might we make the process of understanding scientific literature less intimidating and confusing?
How might we show users how to effectively use the information from scientific literature in their studies and projects?
How might we help users make informed decisions when engaging with scientific literature and conducting research?
03/ Ideate - Creating the Framework
Site Map
Based on the research and persona discovery, I brainstormed, compiled and organized the necessary website content and functionality. Here you can see the navigation will have multiple pages within each category.
Low-Fidelity Wireframes Sketches
In the initial design phase of InSci, I created a set of low-fidelity wireframes to visualize the core structure and functionality of the platform. These wireframes served as a foundational blueprint, allowing for early-stage testing and feedback.
(Originally named SciSimplify as shown in lo-fi wireframes, the project evolved into "InSci" to better capture its comprehensive capabilities. InSci not only simplifies scientific literature but also provides interactive learning and personalized guidance. The name "InSci" echoes the word "insight," representing the clarity and understanding it brings to students, while also standing for "in science," highlighting its focus on making science accessible and engaging.)
Hi-Fi Fidelity Wireframe
The high-fidelity wireframes expanded on these features with a polished, interactive design. They showcased how each feature would function in detail, including new additions such as an interactive chatbot for real-time assistance, and the ability to customize the difficulty level of language used in explanations
04/ Prototype - Where Research and Design Meet
Hi-Fidelity Prototype
I collaborated with Mark Zachary and Sharifa Shari from the MIT Media Lab to create a functional prototype using Github. This first version focused on integrating core features essential for user testing, such as the dashboard, PDF upload functionality, and simplified scientific explanations. Our primary goal was to gather user feedback on these fundamental aspects to ensure they met the needs of high school students. By starting with these essential features, we aimed to establish a solid foundation for the tool. Future iterations will incorporate more advanced functionalities, including an interactive chatbot, customizable difficulty levels for language, and additional research assistance tools, based on insights gained from early user testing.
First Version of Prototype
In the second version of the prototype, we implemented design enhancements based on user feedback and testing insights. We transitioned the website's background color from black to white to create a more vivid and relaxing visual experience, enhancing user comfort during extended use. Additionally, we relocated the document display from a sidebar to within the chat box to better align with users' habits, making interactions more intuitive and streamlined.
(Some features are still in progress due to the time constraints of the project. Key functionalities such as advanced visual aids, which will include detailed diagrams and interactive animations, and adaptive learning capabilities designed to tailor the educational experience to individual student needs, are slated for future implementation.)
Second Version of Prototype
I created a video demonstration (Poke, an extraordinarily intelligent monkey, overcomes the challenges of scientific research with the help of InSci) showcasing the tool's functionalities to first-time users in an interesting way.
05/ Test - To Get Even More Insight and Feedback
Interview Style
1:1 Interviews conducted via in-persons, phone, or Facetime.
Participants
5 high school students between the ages of 16 and 19
Test Goal:
Assess InSci’s ability to help users comprehend scientific literature.
Identify usability issues and areas for improvement.
Gather feedback on educational value and overall user experience.
Methodology:
Pre-Session:
Participants selected a scientific paper of their interest.
They were informed that InSci aims to help high school students understand scientific literature using a Generative AI-based tutor.
In-Session:
Participants were given an overview of InSci, including where to process the paper and ask questions.
They summarized the paper’s findings and asked one additional question.
Post-Session:
Participants compared the paper summary provided by InSci with their own understanding.
They provided free-form feedback on usability, learning, and suggested additional features.
Preliminary Results:
Example User Session:
Pre-Session: User chose a paper on Thin Membrane Disease.
In-Session: User successfully uploaded the PDF and asked questions without needing further instruction.
Post-Session Feedback:
Interface:
(Critical) Switching between asking questions and viewing the PDF was cumbersome.
(Critical) Scrolling up after reading a response to ask another question was inefficient.
Learning Experience:
(Positive) Participant was satisfied with the AI-generated answers, verified as correct through the PDF.
(Critical) Participant was unsure where the answers originated in the PDF. Suggested adding a feature to highlight information used for answers. (Future direction)
(Critical) The tool acted more as an "answering agent" than a "teaching agent." (Addressed by adding AI-generated questions in update)
Wins:
Users successfully used the tool to summarize and ask questions about scientific papers.
Positive feedback on the accuracy of AI-generated answers.
Enhanced confidence in understanding complex scientific concepts.
Opportunities:
Streamline the interface for better navigation between questions and PDF viewing.
Enhance transparency by highlighting the source of answers within the PDF.
Transition the tool from an "answering agent" to a "teaching agent" by incorporating more interactive teaching elements.
Future Directions:
Inference Transparency: Implement a feature to highlight where the AI used information from the PDF, aiding in fact-checking and transparency.
Teaching Agent: Enhance the tool to function more as a teaching agent by incorporating AI-generated and researcher-like questions.
Adaptive Learning: Develop an assessment feature to gauge the student’s comprehension level, allowing the AI to provide more tailored responses.
Progress Tracker: Introduce a visual roadmap to track student progress and promote a sense of accomplishment.
Visual Aids: Enhance understanding with diagrams, charts, and animations.