11/5 Hypothesis Testing

Question 1

Social perception and interaction involve hypothesis testing how?

Answer 1

Candidate for job, student for course, etc.

Attempt to acquire info to test hypothesis.

Question 2

Snyder and Swan. What experiment?

Answer 2

Test hypothesis: introvert or extrovert. Beginning of experiment, might read profile of introvert and extrovert.

To test hypothesis, people chose a series of questions to ask person you are interviewing to test the hypothesis. Choose 12/26 questions. Some oriented to one or another others are totally neutral.

Slides have the example of questions. Questions are biased “What things do you **DISLIKE **about loud parties?” begin with assumptions “When do you seek out new friends?”

**We pick the questions for what we are investigating. **

Many were critical of types of questions offered to participants. Trope and Liberman crticised them

Question 3

Trope and Liberman criticism of hypothesis testing experiment.

Trope is “Mr. Diagnosticity”: why? How? What did he do?

Answer 3

Diagnosticity: the extent to which the subjective likelihood of obtaining some evidence differs if the hypothesis is true from when the hypothesis is false

Greater the difference, greater diagnosticity of evidence. P(X) Given (|) hypothesis true vs. alternative.

he argued that the Swam and Synder questions did not even allow for diagnosticity.

Question 4

How can we determine diagnosticity?

Answer 4

Question Operationalization: how well would the answer to this question discriminate between introvert and extravert?

Sample Introvert:

High Diagnostic: doing things alone more than with others? do they dislike going places with new people?

Low diagnostic: do they study on weekends and watch PBS?

Sample Extrovert:

High diagnostic: strike up convos with people they don’t know well?

Low diagnostic: does this person like beer, are they attractive?

Question 5

New experiment adding diagnosticity ratings, but what else do we add?

Answer 5

Questions extraverts are more likely to say yes to.

Extorverts: say yes to parties

Questions introverts more likely to say yes to.

Question 6

Results of looking at testing diagnosticity (Devine Hirt and Gehrke)

Answer 6

People more often chose diagnostic questions, and more questions that have to do with the hypothesis being tested too.

BUT
People also show bias for response to question they are going to get a “yes” to.

Questions where you’ll get a Yes for what you’re looking for. Confirmation bias, positive test strategy.

Question 7

Aria Kruglanskis’s Theory of Lay Empistemics: motives and process.

What is this? What was he talking about?

**This card: 2 General Epistemic Motivations

Answer 7

1. Need for closure. Some of us want closure or to reach a conclusion quickly. I just need closure, need to have answer. Accentuated by pressure or fatigue or no resources.
2. Need to avoid closure (fear of invalidity): avoid inaccuracy. Accentuated by accountability, evalutaiton apprehension. Don’t want to waste $ on car.
3. Need to seek specific closure
4. Need to avoid specific closure

Don’t want conclusion that you’re about to die or bad person. Favor certain conclusions. Test it to information processing situations.

The difference is: are we trying to avoid closure in general, or a specific conclusion? Sometimes it’s not just a quick or accurate conclusion we’re looking for, it’s a specific one.

Question 8

Illustrative Experiment: Kruglandski and Freund (1983) (Theory of Lay Epistemics)

what does it illustrate?

Answer 8

Illustrates people arriving at conclusion early on a little info or delay until more info.

Ps beleive in study of personnel (employee) selection. Individual head of company, likely candidate for president of company. You test hpyothesis that this person will be successful as company president given how they have been as department head.

Listen to oral recording with + and - behaviors in department head: interst in employee welfare, sensitive to client needs, disorganized, unpersuasive.

Key variable: description is ORGANIZED. Sequence manipulation: + first or - first?

Do people give more weight to what comes first? See difference based on privacy effect. **According to Glancy’s theory, Should be more likely to see primacy effect under certain conditions. **

IVs:
1. manipulate evaluation apprehension (need to avoid closure). Told participants that at the end, participate in discussion where you justify your conlusion to the others in the group. If wrong, you look dumb.
2. Time pressure (need for closure). 2x2 design, put under time pressure.

Likelihood of succeeding at a job on 10 point scale.

DV: data coding to which judgment reflect primacy effect. Higher numbers in positive condition that you have positive judgemetn. Negative condition first, ran with negative to make negative judgement.

Higher numbers = went with info presented first, more primacy effect.

Results:
3 examples of primacy, exception is low time pressure, but evaluation is high. High time pressure: conclude quickly. Low time pressure: make slower decision, AND if you’re worried about how others will see you, you evaluate more information. Avoid primacy effect when you are highly motivated and have plenty of time. Reflects high need to avoid closure (in general, don’t want peers to judge) and low need for closure (specifically, don’t have time pressure).

Depends on balance of these two factors. Remember what they are (a decision vs. specific decision).

Question 9

Webster and Kruglanski (1994)

Need for Closure Scale

Answer 9

Used to think of closure an individual difference.

40 items, 5 factors, controversial, but useful.

Need for closure scale (NFC): referes to desire of an answer, any answer, compared to confusion and ambiguity.

Sample item: “I don’t like uncertainty” “I feel irriated when one person disagrees with what everyone else in the group believes” “I want solutions quickly”

Question 10

Okdie, Rempala, and Garvey (2016)

Need for closure and political ads.

Answer 10

Sample of politically unaffiliated.

• what party do you most closely affiliate? (none, indpendent, or unaffiliates)

Expose to youtube political ads. 2010 US senate race in Colorado. No clear indication of candidate’s political party. Found ads where party not mentioned. No indication of party.

Watch ads in given order. Do people like ad they saw first more? Yes, IF they have a high need for closure.

Krug’s theory gives us a way to think about it. You want decision, go with the first thing you see.

Question 11

Trope and Liberman (1996)

2x2 experiment, Doesn’t focus on motives, uses different language, ends up in same place as last experiment.

How do they do it?

Answer 11

Actual state of affairs: Hypothesis true or false

Decision: Accept or Reject

Hypothesis True/accept = correct acceptance. (also correct rejection)

If true and reject: false rejection (omission)
If false and accept: false acceptance (commission)

We’ve all see the hypothesis table (type 1 and 2).

Decisions about acceptance always have to involve some thershold of confidence.

Reach certain level in information acquired before you know enough to say you can accept or reject the hypothesis.
In different situations, our willingness to falsely accept or reject hypothesis might not be equal.

Question 12

Trope and Liberman’s Model:

The hypothesis table, but what is the point of it?

Answer 12

how long do I continue acquiring information?

Cost of errors vs. cost of information.

Sometimes high cost of information acquisition: need conclusion quick, accentuated by time pressure, fatigue. Then we aren’t concern about errors, want quick conclusion.

High costs of error: Desire to avoid inaccuracy. Accentuated by accountability, evaluation apprehension.

Sometimes it’s not just a quick or accurate conclusion we’re looking for, it’s a specific one.

False rejection costs more than false acceptance, OR does false acceptance cost more than false rejection?

Question 13

Value of Error Reduction

Answer 13

Measure used in Trope and Liberman’s cost ratio analysis.

AKA value of confidence increase.

What is the value of acquiring addtional piece of information to reduce error liklihood? Value of confidence.

Value of that opposed to cost of acquiring extra information.

Futher hypothesis testing is function of Value/Cost Ratio.

Question 14

What are the 3 measures used in Trope and Liberman’s Cost of erros and information model?

Answer 14

1. Value of Error Reduction
2. Cost of Unit of Information
3. Value/Cost Ratio

Question 15

According to Trope and Liberman, what determines whether we will engage in futher hypothesis testing?

Answer 15

Value/Cost ratio

Question 16

Kruglanski and Trope and Liberman: example how they are the same, just different words

K (T and L)

Answer 16

Need for closure (high cost of info)

Fear of invalidity (high cost of error)

Need to seek specific info (false rejection costs more)

Need to avoid specific info (false acceptance more costly)

Go for more info when you haven’t reached content that you are confident.

Question 17

Kruglanski and Mayseless (1988)

Mean Number of Questions Selection: we’re going back to this one. What were Ps told?

Answer 17

Reserach on ability to identify occupation from personality evidence.

Are they a painter? Possible interviewee could be in a different occupation. For example, Could be architect. Now I’ve illuded to possibility could be architect.

Manipulate field and validity.

1. Told this is a very important ability (they want to be good at it). Will have to determine painter based on answers to their selected questions. If successful, get an extra hour of experimental credit for psychology class. Also manipulate need for closure. Emphasis on value of information.
2. Also manipulate time pressure. Ability to make quick decisions about occupations from personality.

Questions they manipulated: diagnostic or not (painters vs. archetect, some questions not).

Table is a mess, just look at his figures.

High diagnostic vs. low diagnostic.

When you fear invalidity (want RIGHT answer) (I don’t want to be wrong, I want that credit!) you chose more high diagnostic questions. In low fear invalidity, it’s about 50/50.

Motivation = seek diagnostic information.

Desire for question that confirms painter hypothesis:

When you Want PAINTER answer: choose more painter related questions, regardless of whether you fear invalidity or not. Positive strategy, want painter evidence.

We got both effects diagnosticity and yes answers.

Question 18

Ditto and Lopez 1992:

What if the hypothesis is about Ourselves?

Answer 18

Need for closure/ undesirability of false acceptance: there are things I don’t want to believe about myself.

Manipulation:
Strip never changes color, what changes is what it means.

Deficiency Condition: if TAA absent, no color reation, if TAA present, changes from yellow to dark green (lack or have enzyme). In this condition, everyone learns they are not okay.

No Deficiency condtion: if TAA present, no color reaction, if TAA absent, color changes to green. In this condition, everyone learns they are okay.

Very defensive when they have been told they are deficent. The Deficiency group says that the serious of TAA deficiency and pancreatic disease is lower and the prevalence of both is lower, and the test isn’t as accurate.

Also told strip may take a certain number of seconds to appear. When complete, put in envelope.

Deficient group waits for longer to accept results and decides to do multiple testing.

You see people want more information.

Question 19

Ditto et al 1998

Mean perceived test accuracy.

Answer 19

Manipulate if having enzyme indicates + or - outcomes this time.

Given information when taking test about error rates. One condition says that some inaccurate readings caused by unusually high blood pressure or blood sugar levels. 1/10 + reactions caused by unusual blood sugar levels. 10% is pretty bad error rates. Vs. only 1/200 reactions caused by blood sugar levels.

Judge test accuracy.

Favorable diagnosis: say test is more accurate regardless of error rate, even if test error rate is bad.

Unfavorable: say test less accurate in the 10% condition, which makes sense, because that’s a bad rate.

Question 20

Epley & Whitchurch (2008)

Self-enhancement

Answer 20

Do I think I’m more attractive than I am (yes, unfortunately)

Ps photos morphed to be more attractive or unatttractive.

2-4 weeks later. 11 photos. Their photos and 10 morphs. Pick which one is you. Random line up.

In all 3, tendecy to think more attractive.

Exp 1 (2-4 weeks later): 10% attractive highest category. Second highest are actual and 10% unattractive. Range is 10% ugly to 30% pretty.

Exp. 1 Follow up (used mirror images to make photos): Mostly 30%, some 10% ugly or even 20%, range even spread. 20 unattractive to 40% attractive. We are more familar with mirror image than actual image. We like mirror, our friends prefer actual image.

Exp 2: piggy backed on experiment where friend accompanies them to the lab. Photos of participant and friend. Photos also taken of experimentor in first phase of experiment and a second experimenter and they too were morphed. In this experiment, rated self as more attractive, friend as more attractive, but not as much, and female and male strangers as less attractive.

Question 21

Smith, Trivers, and VON HIPPEL (2017)

Self-deception and interpersonal persuasion

Answer 21

Because of wishes, we engage in information processing that is self-decpetive.

Goal is to persuade others that mark is likeable or dislikable (or control)

Easier to persuade, if it’s something you believe. If goal is to persuade that they are likeable, it would be useful to believe he’s likeable.

Control condition: own impression, argue for whatever they have.

Watch videos: Mark helps someone, steals money, makes lunch. These are the +, -, and control videos.

Videos in six blocks.
Positive first: first 2 blocks: 2 + , 1 neutral. Second 2 blocks: 1 +, 1 neutral, 1 -, third block: 2 -, 1 neutral.
Negative first is in reverse.

Participants also have different goals.

DV: number of blocks watched before deciding to write speech about Mark..

Control: watched same number of videos for impression regardless of video order.

Told to make him likeable: watched more videos in negative to positive condition, because they needed goal confirming info.

Told to make him dislikeable: watched more videos in positive condition to negative, needed goal confirming information.

Question 22

Hypothesis testing: you see it one the test, what two things are involved?

Answer 22

Motivation and Process

What am I motivated to see?

What processes are involved in this?

Question 23

Readings: Social Identity Theory

Answer 23

Important aspects of self esteem are taken from the groups with which you identify

motivations guide us in what characteristics we think are important for success. Example: if I’m the oldest in my family, and I want to be a doctor, being the oldest is important in becoming a doctor

Question 24

Readings: in group bias

Answer 24

Tendency to see us vs. them even without evidence

Question 25

Readings: Ditto and Lopez 1992 Argue asymmetry in extent of processing

Answer 25

We process and criticize info we disagree with more but accept what fits

Question 26

Readings: What were the two effects of self-serving motivations? What was a third importnat point?

Answer 26

1. Bias how we generate explanations for and evaluate info relevant to our personal interest
2. influence extent we critical evaluate evidence on which feedback with self-relevant implications is or is not accepted
3. Do people ignore info they disgree with? No, actually they contribute considerable attention to it.

Question 27

How does an individuals current motivations and goal change the way they process information?

Answer 27

Bias for explanations that have to do with personal interest

Think TAA enzyme experiment and time it took to see a + or - test.
Think evaluation of Mark: is the goal to make him likeable or unlikeable?

Question 28

How might motives in this part effect (1) inferences and attributions, (2) representation and memory, and (3) impression formation?

Answer 28

1. Atrribution: fundamental attribution error. You might say that someone’s behavior is more indicative of who they are when you want to believe bad things about them. Inference: give insufficient weight to dispositional forces. Ambiguous evalution of someone, we might look at the situation, but not if we want to not like them.
2. You may misremember things, reconstructive memory. For example, Jack’s grades. If you expect him to do better, you remember him doing worse before. Conclude hypothesis about yourself is correct when it is not.
3. Impression Formation: if you are motivated to think gender stereotypes are true, you might be biased towards that information vs. if you don’t want to believe that about yourself. (Category-based impression formation).

Question 29

Readings: Dartmouth Princton football game

Answer 29

Princeton though game was rough and dirty, charges are fair, saw Dartmouth max 2x infractions, and they were more fragert too.

Dartmouth, most said dirty but 1/10+ thought clean and fair. 1/3 said “rough and fair”. Many thought both sides to blame. Thought charges not true. Same number of infractions.

Question 30

Readings: when does occurance become an event?

Answer 30

When it is signfiicant

Question 31

Readings: Main conclusions on self-serving bias?

Answer 31

We see own attributes as predicting more desirable outcomes. Recluctant to accept theories that relate own attributes to undesirable events.

Example: 50% divorce, but not me, because… (lists attributes).

People have optimistic belief that good things will happend to them because of these theories.

These biases do not occur outside motivational pressures.

Question 32

Readings: list self-serving processes

Answer 32

1. Generate causal theories linking any attribute to any outcome “I rest, I perform better, I practiced, I perform better” - divorce is example
2. Evaluate more harshly when it doesn’t confirm what I want. Example: see successful doctor, you want to be a doctor, what traits match me? Well, that must be why he’s successful.
3. Caffeine. Not as convinced that caffine dangerous if you drink it a lot.
4. Reducing motivational pressures to doubt undesired evidence. When the negative caffine evidence was milder, people believed it.

Question 33

Readings: 3 experiments showed that infomation consistent with preferred conclusion examined less critically than inconsistent.

Answer 33

1. Study 1: which of these two students more intelligent? Took longer when the smart one was unlikable
2. Study 2 and 3: given unfavorable TAA test took longer to decided test complete.