Lecture 1 - How to Study + Experiments

Misconception of performance and learning

• “Better performance” = more learning

• Current performance can be a highly unreliable index of whether learning has occurred

Learning

Long-term, improvement that lasts, visible in ANY context.

The permanent change in knowledge or understanding that is the target of instruction - something we must try to infer.

• Learning continued across trials during which the build-up fatigue suppressed performance

Performance

Immediate improvement, visible in a practice context

What you observe during instruction or training.

• Substantial improvements in performance across practice or training sessions can occur without significant learning

Maximizing learning

(Desirable Difficulties) Methods that are most effective for long-term learning tend to introduce difficulties that make short-term performance worse.

Storage Strength

• How entrenched or inter-associated a memory representation is with related knowledge and skills

Retrieval Strength

• Reflects the current activation/accessibility of storage strength and is heavily influenced by factors such as situational cues and recency of exposure.

Relationship of Performance to Storage Strength and Retrieval Strength

• Assume current performance is a function of current retrieval strength, but storage strength acts to retard the loss (forgetting) and enhance the gain (relearning) of retrieval strength

• Conditions that increase retrieval strength the quickest are different from conditions that maximize the increase in storage strength

• If learners interpret their current retrieval strength as storage strength, they may prefer poorer conditions to better conditions of learning

Good strats / conditions to increase learning

• Vary conditions of learning instead of keeping them constant and predictable

• Interleaving instruction separate topics rather than grouping them together by topic (don’t do blocking)

• Spacing instead of massing study sessions on a given topic

• Test yourself

VARY, INTERLEAVE, SPACE, TEST

Vist

Desirable Difficulties

Desirable because they trigger encoding and retrieval processes that support learning, comprehension, and remembering.

• If learning doesn’t have background knowledge or skills to respond to them, they become undesirable difficulties.

Varying conditions of practice

• Constrained conditions = predictable and makes learning contextualized, making it harder for information to be recalled in different contexts / environments.

• Benefits of variation

Spacing

• Massing (opposite) - cramming, short-term performance

  • Rapid gains in apparent learning, but info will be less recallable over time.

• Spacing effect - benefits of spacing on long-term retention

  • Good exam performance and long-term retention

Random vs Blocking + experiment

• Blocking = can make people may perform better in more immediate performance (apparent learning), but doesn’t help with long-term recall

  • Blocked practice appears optimal for learning, but interleaved practice actually results in superior long-term retention and transfer of skills.

• Other benefits of interleaved learning - enhances inductive learning and performances on delayed tests

• Simon and Bjork Experiment

Why might interleaving enhance long-term retention and transfer? (theories)

1. Forces people to notice similarities and differences among different things, which = better encoding = more retention and transfer

2. Interleaving forces learners to reload memories

*Forgetting (losing retrieval strength) creates the opportunity for increasing the storage strength of to-be-learned info or skills

• When some skill/knowledge is maximally accessible from memory, little or no learning results from additional instruction or practice

Generation effect

Long-term benefit of generating an answer, solution, or procedure vs being presented the answer, solution, or procedure.

• Retrieval = powerful “memory modifier”

• Tests > rereading

Pedagogical benefits of tests

• Better recall in long term

• Tests have metacognitive benefits in terms of identifying whether info has or has not been understood or learned

• Tests can enhance the effectiveness of later study sessions, even when conditions make it likely that learners may answer incorrectly

Key ideas

• Spend less time restudying and more time testing ourselves

• Learning requires an active process of interpretation - map new things we’re trying to learn on things we already know

• Retrieval acts to modify memory by making info you practice retrieving more likely to be recallable again in the future in different contexts

Sleep and learning

Sleep organizes and consolidates memories from the day.

Successive relearning

• Practicing a task until it’s performed correctly and practicing it again until you can do it right during other spaced practice sessions

Evaluate strategies - criteria

Strategy - highlighting + experiments

• Fowler and Barker

  • 3 groups: 1 = reads, 2 = reads and highlights, 3 = reads highlighted

  • Performed the same on an exam a week later, Highlighting doesn’t help

• Peterson (underline)

  • 3 groups: 1 = underlines marked article, 2 = underlines blank article, 3 = reads blank article

  • Performed the same on basic factual info exam

  • **Students who underlined did worse on inferences exam

Strategy - issues with highlighting

• May cause us to focus on isolated facts and ignore bigger picture ideas

• Students may struggle to differentiate the main, central idea from peripheral info / extra details

Strategy - rereading + experiment

• Rothkopf

  • Group 1 - never seen passage before

  • Group 2 - read passage once before

  • Group 3 - read passage twice

  • Group 4 - read passage 4 times

Strategy - problems with rereading

• Doesn’t improve comprehension or performance on inference-based questions

• Rereading a 2nd or 3rd time doesn’t help

• Can trick students into thinking that they’ve mastered the material

Recall vs Recognition

Strategy - generating explanations

• Involves coming up with explanations for why concepts are true or how they relate to what you already know

Strategy - generating explanations + experiment

• Presley et. al

  • Groups who made their own explanations performed the best, about 2x as well as the given explanation and no explanation groups

Strategy - interleaved vs blocked practice

• Blocked = study 1 topic until mastery, then move until the next, etc

• Interleaved = mix up problems, jump back and forth from topics

Strategy - interleaved practice + experiment

• Roher & Taylor

  • People who did blocked method did better in practice than interleaved, but those who did interleaved method performed better on exams

Strategy - distributed practice (SPACING) + experiment

• Spread out study over time with breaks

• Bahrick experiment

Strategy - testing

• One of the most effective methods

• Butler experiment

  • Study passage:

    • Group 1 - re-study half the passages

    • Group 2 - test on half the passage

    • Then, both test on the passages after a week

• Strats - don’t just reread; friend can quiz, make flashcards, use Cornell note-taking system

Study strats summary

LOW EFFECTIVE: Highlighting and Rereading

MEDIUM EFFECTIVE: Questions and Interleaving practice

HIGH EFFECTIVE: Distributed practice and Testing

Simon and Bjork

• RANDOM vs blocking on keyboard patterns

• PREDICTION

  • Random group would have more errors than Blocked groups

• RELAITY

  • Random made more mistakes at first, but had almost equal performance as Blocked on 6th practice session

  • RANDOM MADE WAY LESS MISTAKES WHEN TESTED LATER

Fowler and Barker

• Test effect of highlighting on performance

  • G1 - reads

  • G2 - reads and highlights

  • G3 - reads highlighted

• RESULTS: Highlighting didn’t help

Peterson

• Underlining

• G1 underline marked article

• G2 underline unmarked article

• G3 doesn’t underline

• RESULTS 1:

  • Basic factual material test - all groups performed the same

• RESULTS 2:

  • On questions that required drawing inferences, G1 underlined marked did the worst of all

Rothkopf

• Rereading

  • Read once, Read twice, Read three times, Read 4 times

  • Small improvement between reading once versus reading twice (rereading), but reading more than 3 times shows little to no improvement

Presley Et Al.

• Generating explanations

  • G1 no explanation, G2 given explanation, G3 generated their OWN explanation

• RESULTS

  • G1 No and G2 Given performed about the same, G3 those who made their own explanation did far better

Roher and Taylor

• Interleaving

  • Blocked group - studied formulas then did problems, move on to next one

  • Interleaved - read all formulas and interleaved problems

• RESULTS

  • Blocked did far better than Interleaved in practice

  • Interleaved did far better than Blocked in exams

Bahrick Experiment

• Spacing study with breaks

  • G1 - 6 study sessions back to back in one day

  • G2 - 6 study sessions one day apart

  • G3 - 6 studies sessions a month apart

• RESULTS

  • G1 0 Days did best earlier on, tied with G2 1 Day in later study sessions. G3 30 Days in between did worst in all study sessions

• RESULTS 2:

  • G3 30 Days did best on exams, then G2 1 Day, then G1 0 days

  • SPACING = BETTER LONG - TERM LEARNING

Butler

• Testing yourself

  • G1: Read text, test on half of the passages → test on all the passages after a week

  • G2: Read text, re-study half of the passages —> Test on all passages after a week

• RESULTS

  • G1 Practice test group did better on everything