PSYCH 136A Lecture Notes for Midterm

subject expectancies

(try to be good subject and want to please them-conscious or unconscious and consistent to hypothesis)

Subject reactance

(try to do opposite and rebel against study because we dont like being told what to do); we want to show we have free will

advantages and disadvantages of correlational studies?

+ = easy, quick, less $$$. - = you cannot make causation statements. you first need to support a relationship then you can make a true experiment to try to show causation. only true experiments can do this

(challenges existing theories)- The bobo doll study:

- EXAMPLE 1: Bobo Doll study (Bandura): kids see their a video of adults hitting a bobo doll or they see a cartoon hitting the bobo doll. then we see if kids later on hit the doll (which they do). it shows conformity. they learn through modeling. they learn via observation. learning does not need to be done through reinforcement.

differences between random assignment and random sampling?

- assignment is meant for if you want people in experimental group to be similar to control group, you do this - sampling is if you want people in the sample to be similar to people in the real world, then use this - YOU FIRST NEED TO DO THE STEPS IN RANDOM SAMPLING IN ORDER TO DO THE STEPS IN RANDOM ASSIGNMENT

How to show causality

- cannot infer causality from correlational design, only true experimental design - to show causality, you need to hold everything constant except that of IV, because IV is manipulated - to hold everything constant, you need to randomly assign to conditions - generalize results means you need sample from population of interests

systematic error (the more serious one)

- error caused by extraneous variables that tend to influences all scores in ONE CONDITION and to have no effect, or a different effect, on scores in other conditions; so there is error in one condition here. -can distort the effect of IV - hurts I. Validity! - we want to eliminate it all ASAP

true experiment (what does it entail)?

- first you need to randomly assign participants to 1 of 2 groups: experimental or control group! - from there, you manipulate the IV - Then you hold everything else constant besides the IV -measure the DV

how to use true experiment plan with video games and aggression example??

- first you need to randomly assign participants to 1 of 2 groups: experimental or control group! (video games or NO video games) - from there, you manipulate the IV (where or not you play the games in the experimental group) - Then you hold everything else constant besides the IV -measure the DV (measure aggression)

Threats to Internal Validity: History (1): How to deal with them??

- have a control group

Why do we do research in psychology? (4 ways knowing about the world)

- intuition: your gut feeling - authority: asking someone who would know - observation: testing/measuring - logic: using formal principles of reasoning -scientists agree more with observation while math and philosophers choose logic/reasoning

Threats to Internal Validity: experimental mortality--heterogeneous attrition (5): how to deal with them now?

- minimize drop-outs; compare dropouts from the 2 groups and measure each S's change in smoking from Time1 to Time2; do within subjects

Correlations "r" scale ranges from?

-1.0 to 1.0 (-1 being the weakest and +1 being the strongest)

Why do we do research? (hint: 4 goals of psychology research)

-Describes behavior (what happened?) -Predicts behavior (What varies of that?) - Explains behavior (what causes what?) -Changes behavior (what changes what?)

(Integrate existing theories)- Zajonc social facilitation study:

2 competing theories are that: people will perform better in front of an audience OR people will perform worse in front of an audience. -Zajonc study results that audience will help or hurt depending on if the behavior is learned or not SO: if behavior is learned already, you will perform BETTER with audience (ex: professor teaching content for the past 10 years) if behavior is new, audience will HURT performance (ex: a student asked to teach a lecture content)

example of correlational study? hint french paradox

Drinking more wine causes to longer life/better health. Having a better health causes to people to drink more. Something else causes you to drink more and longer health. E.G. being around people or being more happy

(challenges existing theories)- Impact of rewards study:

EXAMPLE 2: Impact of rewards (Lepper): reinforcement actually made it less desirable. 3 conditions: kids told they are given reward for playing, no reward is mentioned, no reward is mentioned but get surprisingly get one at the end. after some time, the kids who play with the toy the most are the ones who got the UNEXPECTED REWARD; the ones who played with the toy the least was the ones who got the EXPECTED REWARD

Comment: we always aim for internal validity first then see if we can generalize it to outside world.

I left it blank. its just a statement

trade-offs between EV and IV?

I. Validity: well controlled and usually lab studies EV: generalizes to other people and other settings, field studies - increase EV by increasing mundane realism or experimental realism -its hard to have high levels of both

IV versus DV?

IV: it is manipulated DV: measured, relies on the IV

Threats to Internal Validity: History (1)

Specific event outside the study/during course of a study that affects the dependent variable (measured). (under category people change) e.g. the pandemic, natural disaster

Threats to Internal Validity: regression to the mean (3) EXAMPLE

Sports illustrated jinx: when athletes who appear on the cover of magazine, they go on to have less excellent performances in their field. BUT ITS JUST REGRESSION TO THE MEAN. Not all performances stay the same! ex: rolling dice study we did there was regression to the mean! example: if we are high in optimism, they will be lower next time regardless of manipulation

main things to consider about correlational studies?

THEY DO NOT EQUAL causation. causation is only possible for true experiments. 3rd variable/confounding variables can happen in correlational studies. so: X causes Y Y causes X Z causes both X and Y

TRUE OR FALSE: RANDOM ASSIGNMENT IS NEEDED FOR CAUSALITY STATEMENTS?

TRUE

TRUE OR FALSE: you need a bigger sample size to truly rep a population?

TRUE

cant have external validity unless study is internally valid? TRUE/FALSE

TRUE

How to demonstrate causation?

TRUE EXPERIMENT

ALL THREATS TO VALIDITY ARE TYPES OF SYSTEMATIC ERROR, NOT RANDOM ERROR? T/F

TRUE(T)

why did we see the results from the previous study? (think motivation)

There is intrinsic/extrinsic motivation: -intrinsic motivation: they enjoy activity, no external reward, so they think: i do this because i like doing it, so intrinsic motivation (FROM WITHIN! - extrinsic motivation: enjoys activity, reward is present like money, they think i do this because i am paid to/forced to, extrinsic motivation(outside of you that makes you feel this way)

Do we need internal validity to establish causality? YES/NO

YES

Does social facilitation happen in other non-human species?

YES

can you use a pretest and post-test to see regression to the mean? YES OR NO

YES

is testing effects the same as practice effects? YES OR NO

YES

confounding variable/alternative explanation

a factor other than the independent variable that might produce an effect in an experiment (the Z)

random sampling

a sample that fairly represents a population because each member has an equal chance of inclusion.

negative correlation

as one variable (X) increases, the other (Y) decreases. they go in opposite direction.

random assignment

assigning participants to experimental and control conditions by chance, thus minimizing preexisting differences between those assigned to the different groups

random assignment

assigning participants to experimental and control conditions by chance, thus minimizing preexisting differences between those assigned to the different groups. you want to give the the equal chance of being assigned to any conditions of the study

threats to internal validity: experimenter bias/influence (7)

definition: experimenter's expectancies can affect either how they act with participants or what they observe

threats to internal validity: participant reaction biases (expectancies & reactance) ((6))

definition: participants try to behave in ways that are consistent with--or the opposite of--the researcher's hypothesis; when the participant realizes they are being studied, they act differently because of it acts in 2 ways: - subject expectancies -subject reactance

Threats to Internal Validity: testing effects (4)

definition: people tend to do better on a test the second time they take it, and they dont even have to see results of first one

threats to EXTERNAL validity: selection bias (10)

definition: sampling people from an unrepresentative sample

threats to internal validity: confounds (8)

definition: some additional variable (you dont care about) varies systematically along with the thing you manipulated; alternative explanation

threats to EXTERNAL validity: experimental mortality- homogeneous attrition (9)

definition: when a particular type of person drops out of study, regardless of which condition they are in; same type of person is dropping out of both conditions

Threats to Internal Validity: experimental mortality--heterogeneous attrition (5)

definition: when different amounts of people OR different types of people drop out of the two conditions of your experiment e.g usually happens in a long study, they get bored etc. think of incentives to keep them in like $$!

Threats to Internal Validity: regression to the mean (3)

definition: when people have high scores on a measure, they tend to have lower scores next time they are tested (closer to the mean) and vice versa! - this happens because their scores are a combination of true ability + random luck so: score on test = true ability + random luck (random luck can be guessing, amount of sleep, etc.)

threats to internal validity: confounds (8)- how to deal with it??

do a true experiment; measure variables that might confounds; do a different study to rule them out

Random error:

error is caused by extraneous and uncontrolled variables whose average influence on the outcome of an experiment is the same in all conditions/it affects everyone in the same way example: bad lighting in a room across all conditions in body image study - subject variables, extraneous events -does not affect validity of results - but can hide the effect of IV - we want to keep random error as small as possible

Threats to Internal Validity: testing effects (4) EXAMPLE

even people who did not take Princeton Review prep course, they tend to do better the second time when taking the SATs

5) speaks to issue of societal importance (example studies)

examples: - stereotype threat study: if asked to identify female before start of math test, they were more likely to do not so great because of stereotype that women are not as good in math -jigsaw classroom (Elliot Aronson): have kids from mixed race into groups to do tasks. after they showed increased work effort, cooperation, mixed racial relationships, and academic achievement

threats to internal validity: participant reaction biases (expectancies & reactance) ((6))-how to deal with them?

give false hypothesis; Convince P hes an E and keep P unaware hes in a sudy; use behavioral measures and not questionnaires; have a non-obvious hypothesis

Threats to Internal Validity: Maturation (2) How to deal with them??

have a control group

Threats to Internal Validity: testing effects (4) How to deal with them?

have a control group; do not give a pre-test; use a different type of test the 2nd time

Threats to Internal Validity: regression to the mean (3)-How to deal with them?

have control group; dont select based on extreme outcomes

threats to internal validity: experimenter bias/influence (7): example

if Experimenter thinks participant has been given alcohol, E would be more likely to notice if P slurs speech

what is the relationship between reliability and regression to the mean?

if MORE reliable/consistent, we will see LESS regression to the mean (so less random luck). and vice versa.

threats to EXTERNAL validity: experimental mortality- homogeneous attrition (9) example

if all the fat people drop out of diet study then you cant generalize the results to fat people

Threats to Internal Validity: History (1) EXAMPLE STUDY

if looking for effects of scary movies on kids' fear of swimming e.g. seeing Jaws , they might be more fearful, but they were fearful not because of scary movies but due to the tsunami that happened in southern Asia in 2004/2006

threats to internal validity: participant reaction biases (expectancies & reactance) ((6)): EXAMPLE

if participants are told they have been given alcohol, they act drunk even if it is not true (subject expectancies)

SELECTION BIAS

in an experiment, unintended differences between the participants in different groups. you chose people on a biased status

Correlation Fallacy

incorrect belief that correlation implies causation

threats to internal validity: experimenter bias/influence (7): how to deal with this?

keep E blind to hypothesis; Keep E blind to what condition P is in; have 2 E's that are each partly blind; have all instructions on tape or laptop

Threats to Internal Validity: Maturation (2) STUDY EXAMPLE

looking at the effects of support group on subjects coping responses, but they may learn to cope better over time even without needing an intervention so it looks like better copers when following up with them at the end

3) challenge people's intuitions or expectations

meant to challenge their expectations or actions

threats to EXTERNAL validity: experimental mortality- homogeneous attrition (9) how to deal with them?

minimize drop outs; compare dropouts to people who remain in the study

(challenge people's intuitions or expectations)-good samaritan study

only 40% of them helped the man on the street. they were on their way to give a lecture so.. if told they had a specific time and were late: the men were in a hurry and did not mainly help the civilian on the street if told they can come in at any time and no hurry: the men were more likely to stop and help the man on the street SO SITUATIONAL FACTORS MATTER MORE THAN PERSONALITY

CORRELATION TYPES

positive, negative, no correlation. positive: r > 0 negative: r < 0 no correlation: r - 0. when the points on graph are all splattered or in a parabolic; this is because correlations happen in linear form.

what are the two kinds of errors?

random error and systematic error

experimental mortality is ONLY A THREAT to internal validity when its heterogeneous attrition (hetero- means different).

so we only talk about different types of people dropping out of the conditions (experimental or control)

Threats to Internal Validity: experimental mortality--heterogeneous attrition (5) EXAMPLE STUDY

stanford three city study: designed to prevent heart disease by making people try to quit smoking! - intervention program happened but did it really work? - who dropped out? 22 people in intervention. why? less motivated, health complications, they are getting annoyed to be told to stop smoking. i just lost lots of heavy smokers so my average went down and the people are different now

SELF-FULFILLING PROPHECY

step 1: perceiver forms expectations of the target step 2: perceiver acts towards the target based on expectations step 3: target intercepts the perceiver's actions and responds to their actions/behavior that is consistent to the perceiver's expectations --its an never-ending cycle

Who has stronger + correlation?

stronger it is, the more linear it is. the more points spread out, then the weaker it is

(challenge people's intuitions or expectations)-Milgram study

student, teacher, experimenter. the STUDENT is actually a confederate. the TEACHER is the only subject here. -if student answers questions wrong, they get shocks. so eventually he complains about the shocks til he remains silent. the experimenter tells teacher to keep going and shocking for incorrect responses. Majority kept going to the max shocks even if there was silence. the few who wanted to stop did stop, but they still reached the max shock. SO CONFORMITY AND AUTHORITY MATTERS

threats to internal validity: confounds (8) EXAMPLE

study from Aronson and Mills tested prediction: if person goes through stressful initiation to get into group, the liking for group increases ---women in sex discussion group and went through initiation in front of men. -no initiation, mild, severe (this one liked the group most)

(results in large effect from a small manipulation)-late-bloomers study:

teachers told some of their students are late bloomers at the start of the year and they should be able to improve at the end. -they werent actually late bloomers, just random assignment. so, the supposedly late bloomers were given hella attention, positive feedback, extra study time and at the end of the year, the supposedly late bloomers had 15 points increase of IQ compared to the non-called late bloomers -so late bloomers performed better because of the extra effort given by the teachers THE TELLING TEACHER WAS THE SMALL MANIPULATION

External validity:

the extent to which the results of a specific study can be generalized to other people, places, or times (so in the real world and beyond the study)

similarity between random assignment and random sampling?

they both want to show similarity among the groups (experimental and control//sample and population)

How to design a valid study? (hint: two types of main validity that we look at)

through external and internal validty

experimental realism

try to make it seem like psychologically meaningful as possible even if doesnt look like real world

mundane realism

trying to make lab setting more like real world

Correlational method

two different variables are observed to determine whether there is a relationship between them. used to see the links between then if possible. the measure of strength between them is also associated.

threats to EXTERNAL validity: selection bias (10)-how to deal with it?

use random sampling

4) results in large effect from a small manipulation

usually revolves around observers expectation so one minor change results in big and long lasting effect

Internal validity:

validity to the extent in which there is a valid causal statement that can be made about the effects IV has made on the DV in the study; so, did IV leave effects on the DV?

threats to EXTERNAL validity: selection bias (10) example:

when the 1936 prez election thought landon will win and not roosevelt but they only sampled rich republican people because they had phones and cars which that was how they were communicated with; newspaper subscribers

Masochistic superstition narcissism

when you think you are the one to blame for something like if your team loses, you think it is because you ate something different that day OR if you didnt wear your sports jersey

2) Integrate existing theories

you add onto existing theories to make it stronger

example of correlation: video games and aggression

you measure video games frequency and measure aggression like amount of time being aggressive. correlation was 0.84 (very strong)

non-response bias

Bias introduced into survey results because individuals refuse to participate. or simply they were not asked to be involved and those responses would have been extremely different

what is involved in studying science? (hint: think of perspectives from leading scientists)

-Ivan Pavlov: (he was involved in classical conditioning). he says there needs to be systematic way of doing things like a plan, the person doing it needs to be modest and aware of own ignorance), remembers science demands your whole life -Isaac Newton: one aspect leads to a whole ocean of research -Albert Einstein: there are always struggles in science -B.F. Skinner: if you run into something, drop everything else and study that new thing

Types of manipulation (manipulation is for IV)

-environmental: of the physical setting example: smoke filled room, objective self-awareness -stimulus: vary actual experimental stimuli -social: depends on action of another person in the experimental situation (example: asch conformity with the lines) -instructional: part of the cover story (example: giving different things to people to change up story, they manipulate story for the condition)

Threats to Internal Validity: Maturation (2)

-hard to hold constant, under category people change definition: changes that happen within people as a function of time passing can affect the DV eg. getting older in age, or get more bored, tired, hungry short term

1) Challenge existing theories

-meant to advance the field and use new data to account for a new theory, tossing out the old theory -following examples are meant to challenge the reinforcement theory(that rewards promote the same behavior 24/7)

(challenges existing theories)-infant bonding study:

-monkeys are separated at birth from their mother. -cloth mother or mother with food dispenser. (both are fake) - we want to see if they go to the mother with food more than the mother with cloth to provide comfort - we ASSUME they want the food. BUT most monkeys went to cloth mother for contact comfort and rather do that than be nourished

(results in large effect from a small manipulation)-self-fulfilling prophecy study (attractive photo):

1 guy and 1 girl on the phone call. guy is given either pretty or ugly picture of girl (not even of the participant, its a random photo). girls not given anything. depending on the photo the guy is given, the girl will act a certain way. - if guy is given pretty picture: girl on end of call will be friendly, likable -if guy is given ugly picture: girl on end of call will NOT be friendly, likable ---his expectations are self-fulfilling prophecy THE PHOTO WAS THE SMALL MANIPULATION

What makes a study important?

1) Challenges existing theories 2) Integrates existing theories 3) Challenges people's intuitions/expectations 4) Results in a large effect from a small manipulation 5) Speaks to issues of societal importance

positive correlation

A correlation where as one variable (X) increases, the other (Y) also increases, or as one decreases so does the other. Both variables move in the same direction.

9 WAYS that threaten validity (the ones that threaten internal validity are the only ones that are systematic error) - first 8 threaten I. Validity -last 1 threatens E. Validity

N/A

another form of mortality which is homogeneous attrition meaning the same.

N/A

do we always have great external validity in psychology studies? yes/no and why?

No, because in psychology we always want to focus on studies in the lab so we tend to have higher Internal validity and not so much external.