Chapter 10 Quiz: Elaboration and Elaborative Interrogation

Answer from memory before checking the guide.

1. Describe the Craik and Lockhart (1972) depth of processing framework. What are the three levels, and what does research show about their relationship to memory?

2. What is elaborative interrogation? What two specific questions does it ask for any new piece of information?

3. What does the evidence show about the effectiveness of elaborative interrogation compared to simply reading facts? Include the evidence grade from the chapter.

4. What is the Matthew Effect in learning? How does prior knowledge affect the ability to elaborate?

5. Explain the self-explanation effect. What did Chi et al. find about students who explain each step of a worked example to themselves?

6. What is the generation effect (covered in Chapter 7 and revisited here)? How does it connect to the elaboration principle?

7. Describe the four steps of the Feynman Technique. What is the purpose of each step?

8. What is the critical difference between the Feynman Technique and retrieval practice? When would you use each?

9. David's hiking analogy connected gradient descent to the experience of walking downhill on a foggy mountain. What makes this a good analogy? What are its limitations?

10. You're learning about supply and demand curves in economics. You've read the definition and looked at the graphs, but it still feels abstract. Apply two elaboration techniques from this chapter to deepen your understanding. Be specific about what you would actually do.

11. What is a concept map and how is it different from an outline or a list? What is the most important element to include when connecting concepts on a map?

12. When is elaboration harder to do, and what are two strategies for elaborating effectively even when you have limited prior knowledge in the domain?

Answer Guide

1. Craik and Lockhart proposed that memory strength is determined by the depth at which information is processed. The three levels: (1) Structural processing (shallowest) — attending to visual/physical features (how does this look? how many letters?); (2) Phonemic processing (intermediate) — attending to sound features (what does this rhyme with?); (3) Semantic processing (deepest) — attending to meaning and relationship to other knowledge (what does this mean? how does it connect to what I know?). Research shows that deeper processing produces dramatically better memory, even without explicit intention to memorize. Thousands of replications confirm this.

2. Elaborative interrogation is a technique that forces deeper processing by asking, for each new fact or concept: (1) "Why is this true?" — what mechanism or principle explains this? (2) "How does this connect to what I already know?" — what prior knowledge does it relate to, what other concepts does it link to?

3. [Evidence: Moderate] Elaborative interrogation produces 30-60% better retention than simply reading facts, across multiple studies. The technique is particularly effective when learners have relevant prior knowledge to connect to.

4. The Matthew Effect is the self-reinforcing cycle in learning where prior knowledge makes elaboration easier, which produces better retention, which builds more knowledge, which makes future elaboration easier. The more you know, the easier it is to learn more. Beginning learners have limited prior knowledge and therefore find elaboration harder — the effect is real but can be addressed by using analogies from other familiar domains and by building foundational knowledge through retrieval practice.

5. [Evidence: Moderate] Chi et al. (1989, 1994) found that students who spontaneously explain each step of a worked example to themselves — identifying which principle is being applied, why each step follows from the previous, what would change if conditions were different — learn significantly more than students who merely read through the examples. The self-explanations force deep semantic processing of the reasoning behind each step, not just the steps themselves.

6. The generation effect (Slamecka & Graf, 1978) is the finding that generating information — even incorrectly — before receiving the correct answer produces better memory for that information than simply reading it. It connects to elaboration because the attempt to generate an answer involves semantic processing: you're actively engaging with the concept, trying to reason to an answer from what you know. This active engagement, even when it fails, prepares the mind to receive and integrate the correct information more deeply.

7. Step 1: Choose a concept to learn. Step 2: Explain it in plain language as if teaching it to a 12-year-old — no jargon, concrete examples, simple vocabulary. Step 3: Identify the gaps — places where your explanation got vague, used undefined jargon, or broke down. Step 4: Return to source material specifically to fill the gaps, then re-explain. Purpose: the exercise forces genuine understanding, because simple explanation requires it. The gaps reveal exactly what you don't yet understand.

8. The Feynman Technique is primarily about understanding — building and verifying genuine comprehension. The quality of the explanation matters; vague or jargon-filled explanations that fail to teach are failures. Retrieval practice is primarily about memory — strengthening the ability to recall information. The quality of the explanation matters less than the act of retrieval. Use retrieval practice when building and maintaining memory for facts, definitions, and procedures. Use the Feynman Technique when testing and deepening conceptual understanding.

9. The analogy is good because it captures the essential mechanism of gradient descent accurately: local slope assessment (the gradient), directional movement toward lower error (downhill), iterative process (reassess and step again), inability to see the whole landscape (working with local information only), and reliable convergence toward a low point. Limitations: the analogy is three-dimensional, while gradient descent operates in very high-dimensional spaces. The hiker can feel multiple slope directions simultaneously; gradient descent must calculate the gradient explicitly. The hiker has one trajectory; modern optimization uses variants (mini-batch, momentum) that the analogy doesn't capture. Good analogies are always partially wrong — what matters is that they give you a functional mental model to start from.

10. Many answers acceptable. Example for supply and demand: Elaborative interrogation — Ask "Why is this true?" about the downward-sloping demand curve: Why do people buy less of something when the price rises? Because of substitution (you use alternatives) and income effects (you have less purchasing power). Now ask "How does this connect to what I already know?" — it connects to everyday behavior when your favorite restaurant raises prices. Self-explanation/analogy — When you look at the graph, explain it out loud: "The curve slopes downward because at higher prices, fewer people are willing and able to buy. The intersection of supply and demand — the equilibrium — is the 'negotiated agreement' price where the market clears. It's like two people haggling: the buyer wants the lowest price, the seller wants the highest, and the equilibrium is where both can agree." Then generate a personal example: what has price changes made you buy more or less of recently?

11. A concept map is a visual network diagram showing how concepts relate to each other — not a linear list or hierarchical outline. Unlike an outline (which shows hierarchy) or a list (which shows membership), a concept map can show multiple types of relationships between any two concepts, including cross-connections that don't fit a hierarchical structure. The most important element is labeling the connecting lines with the specific type of relationship: "causes," "is a type of," "requires," "increases," "is the opposite of." Unlabeled lines just show that two concepts are "related somehow," which is shallow. Named relationships force you to articulate exactly how the concepts connect.

12. Elaboration is harder when you have limited prior knowledge in the domain — you don't have much existing knowledge to connect to. Two strategies: (1) Reach across domains — even without domain knowledge, you have knowledge from everyday life, other subjects, and personal experience. Use analogies from familiar domains even if they're imperfect. (2) Use elaborative questions as prompts even without full answers — "Why might this matter?" "What conditions would change this?" — generate the questions even if you can't answer them fully yet. The questions prime you to seek and recognize the answers. Additionally: build foundational knowledge first through retrieval practice and spaced repetition, then elaborate on a richer base.