Preparing for SAT Based on Scientific Research
Whenever I need to research just about anything, I make use of the amazing research-based AI tool called Ai2 Asta from Allen AI. Since my daughter is preparing for the SAT, I decided to look up research around studying and test-taking tactics for the SAT.
The most successful approaches focus on test format training, reasoning and deduction skills, time management, practice exams, and coaching strategies.
Studying/practicing
Start Early
- Students who begin test preparation earlier in their junior year consistently achieve higher scores, as early start timing allows for more distributed tutoring sessions, increased instructional hours, greater homework completion, and more practice test opportunities (Appelrouth et al., 2018).
- The distribution of study sessions over time, rather than intensive cramming periods, shows a positive direct effect on SAT score increases, along with both individual and group tutoring hours and the number of practice and official tests taken (Appelrouth et al., 2018).
Practice Tests Have Significant Impact
- Practice testing emerges as a particularly powerful preparation tool, with students gaining approximately 20 points on official SATs for each simulated SAT trial, and potential gains of 40 points resulting from just four to six hours of practice trials (Appelrouth et al., 2018).
- Students who begin their preparation by taking an unofficial practice test demonstrate significantly larger improvements, gaining 37.25 points for each subsequent official SAT compared to only 20.30 points for students who start with an official SAT as their baseline (Appelrouth et al., 2018).
- Taking the SAT multiple times also proves beneficial, with 64 percent of students achieving their highest single administration score on their final attempt (Appelrouth et al., 2018).
Study Tactics
- Effective SAT preparation programs incorporate specific training components including practice tests, review sessions, and online training programs, with curriculum that contains test-taking strategies showing particular effectiveness (Xu et al., 2016).
- Active rehearsal of study materials emerges as another critical practice, with students who put effort into rehearsing texts and notes performing better on tests compared to their counterparts. This rehearsal should be combined with help*seeking strategies, where students actively seek assistance from teachers, tutors, or peers when they encounter difficult concepts.
- Self-testing is one of the most effective study strategies, with students who use practice testing showing better performance (KeinyatseKgosidialwa et al., 2019). This strategy allows students to identify knowledge gaps and practice recall under test-like conditions.
- Comprehensive note-taking techniques significantly impact test scores, with research showing that the quantity of notes has a direct correlation with higher performance. (Al*Shammari, Zaid N, 2019). Effective note-taking involves condensing original information, extracting key concepts using critical thinking skills, and creating concept maps to reorganize information (Al*Shammari, Zaid N, 2019). Students should also engage in highlighting important information and summarizing content as part of their note*taking process (Al*Shammari, Zaid N, 2019).
- Small group tutoring enhances academic gains due to personalized feedback and adaptive teaching. Programs with more total hours of instruction, especially beyond eight hours, show stronger effects on math performance.
- For intensive vocabulary building, students can use flashcards combined with mnemonic techniques to learn large numbers of words in short timeframes. One approach involves creating flashcards for all unfamiliar words from preparation books and using basic mnemonic strategies, such as connecting new words to familiar concepts or images (Werner, 2020). Vocabulary preparation works best when combined with other engaging study approaches and integrated into a broader, more varied preparation program.
Online Study Platforms
- Use of Online Study Platforms: Common and effective online tools include CrackSAT.net, PrepScholar, and Khan Academy. CrackSAT helps students review specific topics with problem explanations; PrepScholar provides downloadable practice tests; and Khan Academy, in partnership with the College Board, offers personalized study schedules, progress tracking, and daily task boards to guide structured practice
Test Taking Strategies
- Question analysis techniques: Students should underline key words in test questions to better understand what is being asked and focus their attention on the most important elements (Dorji et al., 2025)
Reading and Writing
- Skim the passage: Read through the passage quickly to get a general understanding before looking at the questions.
- Predict your answer: Before looking at the multiple-choice options, try to formulate your own answer in your head. This helps avoid being misled by distractors.
Address question types
- Words in Context: Find the specific line, read the surrounding lines, and think of your own synonym before checking the options.
- Command of Evidence: Find the evidence in the text that supports your answer choice and ensure it directly supports the claim made in the question.
- Inference: Look for logical conclusions based on the information given, but avoid overthinking or making assumptions not supported by the text.
- Skip if stuck: If you can’t answer a question within 30 seconds, skip it and come back later to save time.
Math
- Understand question types: Be familiar with questions asking to use ratios, rates, percentages, and scale drawings.
- Focus on data analysis: For questions involving tables and graphs, read them carefully and understand the relationship between the data points before answering.
- Use estimation: In data interpretation questions, you can often estimate to quickly eliminate unrealistic answer choices.
- Check your logic: Ensure that your chosen answer logically completes the sentence or answers the question in the context of the data presented.
General Strategies
- Be familiar with the test: Look at past tests to understand the layout and types of questions you will encounter.
- Read the question twice: Make sure you understand what is being asked before you start answering.
- Focus on evidence: For reading questions, the correct answer will always be supported by evidence directly from the passage.
- Answer elimination methods: Systematically eliminating obviously wrong answers before making a final selection, which increases the probability of choosing correctly among remaining options. To use the answer elimination method on the SAT, carefully read the question and look for keywords, then eliminate answer choices that are clearly wrong, too extreme, off topic, or contradict the passage. In math, you can “plug in” the answer choices to see which one works. For reading, you can “plug in” the answer choices by substituting the words back into the passage to see which one fits the context and meaning best. If you are still stuck between two answers, compare them sidebyside to find key differences in logic, tone, or detail.
- Strategic question ordering: Answering easier questions first to build confidence and ensure maximum points are secured before tackling more difficult items
- Educated guessing strategies: Using specific techniques such as selecting longer answer options when unsure, looking for similarities or opposites among choices, and applying deduction reasoning to narrow down possibilities (Hamzeh et al., 2015)
- Time management during testing: Allocating appropriate time to different sections and questions to ensure the entire test can be completed within the given timeframe (Dorji et al., 2025)
Other Test Taking Tactics
- Mastering Command of Evidence (Textual) Questions on the Digital SAT
- The New SAT Reading Test: Strategies for Question Types | CollegeVine Blog
Self-Control Strategies
- Students who engage in goal setting and strategic planning during test preparation show significantly higher mathematics performance, as goals help maintain motivation and provide standards for self evaluation (Tee et al., 2018). The planning process should include creating structured study schedules and setting specific learning objectives for each preparation session.
- These self-control strategies include practical techniques such as choosing to study in environments that make it easier to focus and implementing other environmental and behavioral controls to support sustained preparation efforts (Baldwin et al., 2022).
- Students also benefit from incorporating relaxation techniques such as slower breathing during studying and test taking, which can improve information processing and recall when combined with effective study methods (Peper et al., 2016).
Our AI future: Good, Bad, and Maybe Nirvana
1-5 years
- AI, Agentic software, Robotics, Quantum tech, and 3D printing see exponential increases in effectiveness. If truly efficient and capable quantum is realized then this timeline accelerates dramatically
- All 3 start eating eating away most jobs
- Companies race to acquire other companies in their segments to stave off race to bottom on pricing as efficiencies dramatically increase.
- Companies that didn’t go 💯 in AI die off rapidly
- Supply chain begins to shrink
- Lack of income decreases global sales
- Crypto becomes the defacto payment system
- All of compute will be AI
- Apps will die
- The operating system as we know it will no longer exist. Zero need for the overhead
- We begin to enjoy and experience AI as a central part of our lives. It still feels new and strange.
5-10 years
- AI has become common as part of our everyday interactions
- Markets begin to collapse
- Governments attempt UBI and that works for about 5 minutes, but the dramatically reduction in tax revenue from job loss makes it impossible to continue
- Supply chains start to collapse because the gaps are not being filled rapidly enough to make up for the losses of companies that couldn’t automate fast enough.
- More localized & efficient and large scale 3d printing/robotics starts killing mass manufacturing
- Shortages abound until can be compensated for locally with printing and robotics
- UBI no longer works, because no revenue to pay for it and printing $$ loses power. The race to the bottom dynamic makes it worse.
Eventually
- We are now in what I call the Chasm of Chaos
- All markets collapse
- Money loses all value
- No leaders taking the difficult steps for a post capitalism existence
- Lots of chaos and suffering in the streets from food and medicine shortages
- Most likely violence and crazed cults spring up
- Countries will likely invade other countries to secure minerals needed for the tech
- Additionally, there is a chance that AI is lost in the chaos and we have to rebuild it to get to potential nirvana
- Think that humans will figure out a way over the Chasm, but much suffering will ensue.
- On the other side is a nirvana of human existence.
If Nirvana
- Especially with quantum, most disease and illness eradicated.
- Lifespans will be dramatically increased with longevity science driven by AI.
- Body part replacements will be common.
- Humanity will start looking more and more like singularity predictions, but not required.
- Energy solved. Hopefully no one company or country has control over the tech.
- Climate solved
- Creativity/art explodes. Real life performance and creation become incredibly valuable.
- Space travel deep and near becomes normal
- Expect at some point a rotation of humanity for a desire to have real life existence experiences. Back to nature, etc
- Focus on mindfulness is a likely outcome.
- Much more peaceful world with bountiful supplies of everything humans need on demand.
Keys to successfully getting over the Chasm of Chaos
- Suffering is likely required. Nobody is going to take the chance to implement post capitalism existence until suffering.
- Reject centralized AI. If not, then 1 or 2 companies control all of humanity or we lose control of AI and kills us all.
- US govt must not be dependent on Grok/Musk for AI. Must be distributed/decentralized.
- Avoid world war. Critical that resources are shared equitably or 1 or 2 countries will control all. Governments must secure their own & also work to secure others collaboratively
- Humans need to plan now for transition to reduce future suffering. The problem is whether “leaders” will see what’s coming.
- Build up infrastructure globally to support AI.
- Emphasis on personal freedom including with AI. Systems that allow humans to own their AI will prevent centralized control of mind & body.
What are your thoughts? Does it make sense or do you think there is a different path?
What are the hottest🔥 projects in open source AI this week (3/6/2025)?
OpenXAI: Benchmarking AI Explanations for Trust and Transparency Description: OpenXAI is a library for benchmarking explanations for structured data in machine learning models, promoting transparency and reproducibility in AI systems.
Link: open-xai.github.io
RL Swarm: Distributed Reinforcement Learning for Scalable AI Training Description: RL Swarm is a framework for collaborative reinforcement learning over the internet, enabling efficient and scalable distributed training for AI models.
Link: github.com/gensyn-ai…
Verde: Securing AI Models in Decentralized Environments Description: Verde is a verification system for machine learning over untrusted nodes, ensuring the security and integrity of AI models in decentralized settings.
Link: www.gensyn.ai
SkipPipe: Efficient Communication for Scalable ML Training Description: SkipPipe is a communication-efficient method for decentralized ML training, optimizing data sharing to reduce overhead and improve scalability.
Diverse Expert Ensembles: Building Robust LLMs with Diverse Expert Knowledge Description: Diverse Expert Ensembles offers an approach to building large language models using parallel processing and diverse data, improving model robustness and generalization.
Link: www.gensyn.ai
JailbreakMe: Community-Driven AI Security Testing Description: JailbreakMe is an open-source dApp that allows testing AI models for vulnerabilities, enhancing security through community engagement and rewards.
Link: github.com/jailbreak…
Llamacpp: Bringing AI to Your Device with Local Execution Description: Llamacpp is a C++ implementation of the LLaMA model, enabling users to run large language models locally on their devices, enhancing privacy and control.
Link: github.com/ggerganov…
Open WebUI: User-Friendly AI Interface for Everyone Description: Open WebUI is an open-source web interface for AI models, providing a flexible and customizable platform for AI interactions with support for various LLM runners.
Link: openwebui.com
Deforum: AI-Powered Animation for Creative Expression Description: Deforum is an open-source AI animation platform using Stable Diffusion, allowing users to create AI-generated animations from text, fostering creative expression in AI art.
Link: deforum.art
I’m old enough remember when all screens and apps were square.
First micro.blog post using Drummer
Decided to work a bit with Micro.blog and get it setup. My biz partner/friend Chuck Shotton suggested it. That then led me to Drummer by Chuck’s friend and my old acquaintance Dave Winer. So, here we are trying out Drummer connected to Micro.blog. So this will be my first post on Micro.blog, but first I have to figure out how to connect it to micro.blog.