Bruce: Problem solving

Question 1

Define difference reduction (hill climbing)

Answer 1

Difference reduction aims to reduce the difference between the current state and the goal state, by taking one step in the right direction

Question 2

List 2 advantages and 2 disadvantages of difference reduction (use the laundromat example)

Answer 2

Advantages:

• Steps can be small
• You don’t have to know how much, just what gets you closer to the goal (don’t need to know how far away the laundromat is, just which direction it’s in)

Disadvantages:

• Requires some knowledge (eg. where is the laundromat)?
• It may not always be possible (or desirable) to get closer on each step (eg if there’s an earthquake line, you need to walk in the wrong direction first)

Question 3

Should there be an intrusion, rendering difference reduction not possible (such as an earth quake creating a chasm between the current location and destination), what problem-solving method would you need to apply, and why?

Answer 3

I’d need to apply means-end analysis rather than difference reduction if there is an added obstacle, since difference reduction doesn’t allow for constraints. Means-ends analysis allows me to do something else (sub-goals) in order to reach the end.

Question 4

What would be the order of going about solving a problem using means-ends analysis? Think about the tower of Hanoi and identify 4 steps.

Answer 4

Using means-end analysis:

1. compare current state to goal state and identify differences eg DIFFERENCE REDUCTION
2. Select an operator to reduce the LARGEST difference (eg we need to move stone 3 onto post C).
3. If this operator can’t be applied, set a subgoal of creating preconditions for it’s application
4. Return to step 1 (comparing current state to goal-state) and follow the steps through until the the goal is reached.

Question 5

What is the difference between decision making and problem solving?

Answer 5

Whilst problem solving is finding a path to a goal, decision making is choosing between available options

Question 6

How does means-ends analysis affect the difficulty of problem?

Answer 6

Problems become more difficult with more subgoals, and the probability of error is increased the more subgoals are necessary.

Question 7

Anderson (1983) showed that people sometimes set subgoals while problem solving. How did they do this?

On the contrary, Catrambone (1995) showed that people were not using subgoals whilst solving maths problems. How did they show that?

Answer 7

Anderson (1983) had people speak through the process they were using whilst solving problems, and found that some people spontaneously set subgoals.

However, when Catrambone (1995) trained people to use subgoals, their maths performance improved, which suggests that they weren’t previously using subgoals.

Question 8

How did Patsenko & Almann (2010) show that people weren’t using detailed planning whilst performing the Tower of Hanoi challenge?

Answer 8

During the Tower of Hanoi challenge, Patsenko & Altmann (2010) tracked participants eyes and actually changed the number of discs on the towers during saccades. Since the participants didn’t notice, they concluded that they weren’t using much planning.

Question 9

What are the 2 heuristics which people commonly apply to problem solving, flagged by Ormerod et al (2012) as the Criterion of Satisfactory Progress (CSP)?

Use the 9 ball challenge as an example.

Answer 9

CSP (Criterion of Satisfactory Progress) states that when problem solving, people often apply these 2 heuristics:

1. Minimisation of search space - this makes the problem less overwhelming
2. Maximisation of progress (eg put the most balls possible on the scale). This is effectively difference reduction, in that it is a step in the direction but with no subgoals.

Question 10

Why did the Piltdown man scam work? Reasons include confirmation bias, applying simple rules and motivated reasoning. Define the latter 2 giving examples.

How does motivated reasoning follow on from confirmation bias?

Answer 10

The simple rule in this case was ‘trust the experts’. Although some of the people who were fooled were themselves experts, they were swayed by the other experts opinions.

Motivated reasoning was applied given that if Piltdown man was real, it would mean the oldest specimen was now from England rather than Germany where the Neanderthals are from.

Motivated reasoning allows us to argue against non-confirming information (an extra step on from confirmation bias which is a bias towards information which supports our hypothesis).

Question 11

Define and give examples of both inductive and deductive reasoning.

Answer 11

Deductive reasoning takes into account which conclusions would likely follow the given information, assuming that is true.

For example, if Tom is taller than Dick, and Dick is taller than Harry, it follows that Tom is also taller than Harry.

Inductive reasoning is commonly used by scientists and detectives. Conclusions are drawn based on the available evidence. This is not guaranteed to produce valid conclusions.

For example, a turkey assumes that it will be fed tomorrow since it has been fed every other morning, but if tomorrow is Christmas eve, its assumption will be false.

Question 12

“Trust the experts”, “follow the doctors advice” and “memorable things are important” are all examples of _________.

What is the main reason people utilise these, and how do they result in predictable errors?

Answer 12

These are heuristics. People utilise heuristics because it makes difficult decisions easier. Predictable errors happen when the heuristic doesn’t ring true - eg ‘trust the experts’ in the Piltdown man scenario. There wasn’t enough information available yet to disprove it.

Question 13

Representation is critical. How does this apply in the Piltdown man scenario?

Answer 13

It’s harder to recognise something as a hoax when it’s being framed as legitimate - I would need to gather more new evidence against the evidence that’s there.

Question 14

What is the Gestalt approach to problem solving? Describe reproductive and productive thinking. Which is more complex?

Answer 14

The Gestalt approach emphasises looking at the problem from different perspectives.

Reproductive thinking is reusing prior knowledge and thought patterns, whereas productive thinking involves approaching the problem in novel ways. Productive thinking is more complex.

Question 15

What is functional fixity? Think of the candles thumb tacks problem.

Answer 15

Functional fixity is when we become fixated on the functionality of objects, and see things in certain ways, using the same kinds of solutions which we get better and better at.

Question 16

In the candle/ tacks/ wall problem, what did Adamson (1952) find when the box of thumbtacks was empty, as opposed to the other condition where it was completely full?

What was the problem participants had?

Answer 16

Adamson (1952) found that 86% of people solved the problem when the box was empty, but only 41% when it was full.

the problem participants had here was functional fixity.

Question 17

After a sequence of similar solutions, people often fail to notice an easier solution.

What is this effect? Think of the measured jugs problem, and how numerous repetitions are necessary to create this issue.

Answer 17

This is a set effect, when people fail to think of an easier solution. The 1st jugs problem required multiple decanting whereas the last one could be done with only one action, but people often deferred to the previous methods and failed to see the shortcut.

Question 18

How did astronomer Le Verrier fall victim to the set effect and how was he finally proven wrong after his death?

Answer 18

Le Verrier used the same solution which Newton discovered to predict the location of Neptune, but failed numerous times. Finally Einstein came up with the theory of relativity, which explained why the previous formula didn’t hold.

Le Verrier fell victim to the set effect, whilst Einstein adopted novel thinking.

Question 19

Explain how representation is vital for Gestalt theory. What can changing representation lead to?

Answer 19

Gestalt theory poses that problem solving requires novel perspectives; in other words, if we change the way we represent the problem we might encounter insight, or ‘aha’ moments.

Question 20

The Gestalt theory seemed largely descriptive. How did Newell & Simon (1970) expand the definition of problem solving by defining the process?

How did this apply to their project of programming AI, and how does this compare to humans?

Answer 20

Newell and Simon (1970) proposed problem solving as a search through a problem space of posibilites.

In the case of AI, this problem space is a physical space to be navigated (or a software space), whereas in humans it’s a mental space.

Question 21

Newell and Simon’s (1970) point out, as part of their problem space proposition, that search for possible outcomes can’t be exhaustive.

What is their proposed solution to exhaustive search?

Answer 21

Their proposed solution is that we need to apply strategies, since we can’t search exhaustively.

Question 22

An action that changes one state into another is known as an ________.

Answer 22

An action that changes one state into another is known as an operator.

Question 23

A sequence of operators that transforms initial state into goal state is a ________.

Answer 23

A sequence of operators that transforms initial state into goal state is a solution.

Question 24

Generate-test is one approach to problem solving. What is it’s advantage, and how likely is it to be effective? Think about breaking a combination lock.

Answer 24

Generate-test involves random generation of a possible solution and then testing its effectivity. Its advantage is that no prior knowledge is needed, but it’s quite ineffectual.

Question 25

Attentional control, distraction resistance and ability for fine focus are all examples of _________.

Answer 25

Attentional control, distraction resistance and ability for fine focus are all examples of WMC (working memory capacity).

Question 26

What is the effect of WMC on creative problem solving, and why?

Answer 26

Creative problem solving can actually be hindered by WMC, because it requires a broader scope of focus, and WM is too finely tuned.

Question 27

What did De caro et al. (2016) find about correlations between WM and incremental problem solving VS insight?

found that increased correlated with increased ability for what type of problems, as opposed to insight problems? Why is this?

Answer 27

De Caro et al. (2016) found that increased WM correlated with ability to problem solve incrementally but not using insight.

Question 28

If I’m applying an incremental approach to problem solving and reach an impasse, what do I naturally do?

Answer 28

When I reach an impasse, I apply a representational change to the problem which might lead to an insight.

Question 29

There are some interesting effects of insight problem solving compared to analytic problem solving.

What was found during alcohol intoxication, problem-solving at a non-optimal time of day and for bilingual people?

Answer 29

Bilingual people are better at insight problem solving than monolingual people.

Insight problem solving is enhanced at non-optimal times of day, compared to analytic problem solving.

Alcohol intoxication actually increases insight problem solving ability, but not so much analytical problem solving.

Question 30

What is the difference between simultaneous VS sequential eyewitness identification, and which results in more false positives?

Answer 30

In a sequential lineup, the witness views each suspect one at a time. IN a sequential lineup, they view all the suspects together.

Sequential lineups produce more false positives; people feel obliged to pick one of the suspects even though it might not be any of them.

Question 31

What is involved with the problem solving heuristic ‘anchoring’?

Which related heuristic is this aligned with?

Give an example (completed in the tutorial).

Answer 31

Anchoring is using some arbitrary information to guide a decision when the question is too hard.

Anchoring is aligned with the heuristic

For example, I used the last 3 digits of my phone number as an anchor to estimate the date of Atilla the Hun AD.

Question 32

The conjunction fallacy states that the conjunction of 2 events can’t be more likely than one event, what is a another term for this heuristic?

Give an example involving Linda.

Answer 32

Another name for the conjunction fallacy is the representativeness heuristic.

Linda is described as someone interested in social issues. When we’re asked to judge whether she is more likely a bank teller or a feminist bank teller, ,most people think she represents someone who would be a feminist, thus choose the latter, which is numerically less likely.

Question 33

Thinking of trivia time in tutorial 5, when presented with questions with either 2 or 4 response options, what was the effect on confidence that responses were correct?

Why?

Answer 33

The more options are given, the more confidence increases. This may be because the more we reject, the more confident we are in what we’ve selected.

Question 34

People are unlikely to accept a 50/50 bet unless the amount they might win is about twice as much as the amount they might lose.

What is this an example of?

Answer 34

This is an example of loss aversion.

Question 35

Define the availability heuristic in the context of estimates of ‘types of death’.

Answer 35

People refer to their LTMs to access available information, for example ppl estimated murders as a more likely cause of death than suicide because of the available media coverage of murders.

Question 36

In Gick and Holyoak’s (1980) experiment comparing the tumour - rays problem with and without an analogy presented, how many more people reached the correct conclusion given an analogy?

How much did correct response improve if a hint to use the analogy was given?

Answer 36

When given an analogy, 40% of people reached the correct conclusion, compared to 10% without an analogy.

If a hint was given to use the analogy, correct response was almost 80%.

Question 37

Give examples of how transfer of old knowledge onto a new problem can be negative or positive.

Is an analogy a form of transfer?

Answer 37

Negative transfer is fixedness and positive transfer is when that prior knowledge which is transferred actually helps to resolve the problem.

Yes! An analogy is a transfer.