ap psychology: research unit

theory

possible explanation of a phenomenon → helps validate

example - the more you see someone, the more you like them

THEORY AND HYPOTHESIS ARE NOT THE SAME

hypothesis

testable prediction of an outcome (or a study) → helps form

example - subjects will more likely want to play a game with people they saw earlier opposed to people they just met

operationalization(operational definition)

how is it measured?

(is the measure good/fair/makes sense?)

how variables in a study are measured

types of measures

survey, test, inventory

validity

is it accurate?

reliability

is it CONSISTENT?

population

group of people who possess a certain characteristic (who I may want to study)

sample

smaller group that represents the population

the group participating in the study

internal validity

measure assesses what it is supposed to

external validity

sample represents the population → results from sample can be applied to population

random sampling

people from the population who are chosen randomly → good external validity

sample bias

poses the greatest threat to external validity → poor external validity

case study

• poor external validity

• rare concept with much info

• provides the most detailed qualitative information

• allows you to collect a lot of information from something rare or unusual that is likely not possible with other research methods

• a type of a research method that you can’t really conduct repeat and experiments on so you use the all information that is already available

meta analysis

uses previous studies and puts it all together

experiment

• study 1, study 2

• variables that can be manipulated

observational study

• naturalistic → variables you can’t control

• laboratory → artificial

pro and con of experiment

infer cause and effect

can be complex

variable: dependent

• it’s affected/depends on the independent variable

• you can’t control

variable: independent

the variable you can control and change

variable: control/controlled

the variable you keep the same throughout the experiment

variable: extraneous

variables you can’t control

confounding

• interferes with the influence of the independent variable

• a kind of extraneous variable that is messing things up; makes things confusing

• variables that adversely affect an outcome of a study that we cannot control and/or are not aware of

the effect of extraneous/confounding variables

researchers take into account extraneous variables because the whole experiment must be considered invalid if another variable affects the study because the experiment is longer testing one’s variable’s affect on the other

control group

no special treatment

experimental group

gets special treatment

ethics

• IRB approval

• confidentiality

• informed consent

• no coercion

• minimized risks

• deceptions when necessary

• debriefing

IRB (Institutional Review Board)

• approves any kind of research before it starts

• reviews research purpose

• will accept or reject and ask you to revise

• looks at research ethics

• they do NOT fund the study

confidentiality

the participant can tell whoever but the research has to make sure the participant’s identity is kept a secret unless it’s an emergency like they’re going to hurt themselves

informed consent

participant has to give you their consent to do the experiment unless you’re in a public space like a mall

no coercion

researcher cannot force a participant to do anything nor can they offer money

minimized risks

ALL research has risks but it is always better to have less risks

the IRB asks “does the benefits outweigh the risks?” example - coronavirus vaccine

deception when necessary

it is permitted under certain circumstances

debriefing

tell participants about all the aspects of the study (what is going on, the purpose, hypothesis, etc.) following participation in a study

random assignment

each person has an equal chance of being in either the control or experimental group

avoids experimental bias

double blind procedure

neither the researcher nor the participant knows the participant assignment

avoids experimental bias

single blind procedure

the researcher knows the assignment of the participant while the participant does not know their assignment

experimenter bias

• researcher influences results

• sometimes deliberate, sometimes not

social desirability bias

appealing better than one is in a study

examples - rating oneself better, knowing like you have knowledge

placebo effect

“fake pill” → healing response

expectation of healing → healing response

hawthorne effect

knowing you’re being studied which facilitates behavior

correlation

the relationship between two variables

NOT CAUSATION (just because there’s a correlation between two variables does not mean one variable caused the other variable to happen)

spurious correlation

• a misleading/non-existent relationship

• two variables that appear to correlate but in fact do not

positive correlation

same direction: A↑ B↑ and A↓ B↓

POSITIVE AND NEGATIVE DO NOT REFER TO GOOD AND BAD

negative correlation

opposite direction: A↓ B↑ and A↑ B↓ (inverse)

POSITIVE AND NEGATIVE DO NOT REFER TO GOOD AND BAD

correlation coefficient

• direction → positive (+) or negative (-)

• strength of relationship → how much do the dots look like a line? (number from 0 to 1)

0.2

weak correlation, dots are going in a positive up direction

-1.0

perfect correlation, dots made a straight line that is going in down negative direction

mean

• the average

• most easily affected by extreme scores

median

• middle value

• is the middle of all vales

mode

• most frequent mode

• can be more than one

• mode is the highest point

normal distribution

mean=median=mode

positive skew

mean>median>mode

negative skew

mode>median>mean

measures of spread

=measures of variation

range equals

highest-lowest

standard deviation

• standard measure or variability

• can compare one variable to another

standard deviation equals

square root of variance

☆ the higher the SD, the bigger the spread

variance equals

square of standard deviation

☆ the higher the SD, the bigger the spread

EX: 100 mean (0 SD)

1 above: 110

1 below: 90

SD is 10

z-score

number of standard deviations above or below the mean

z score equation

z score = score-mean/sd

if the mean of test scores was 100 and the standard deviation was 15, what is the z-score of a person who got 130?

130-100=30 30/15=2

(consider a normal distribution) given a mean of 50 and a standard deviation of 5, what percentage of participants had a score between 45 and 55?

68%

percentile

what percentage of population did a value outscore? → add up everything to the left of the value

percentage is the same as _____

percentile

suppose a person has a score of 1.0. what is that person’s percentile?

84

what is the percentile of someone who score corresponds to a z-score if -1?

16th percentile (or 15.5 if you don’t round)

concluding an effect or difference exists →

conclusion is wrong →

type 1 error → chance of this → p-value

any p-value equal or less than this is acceptable

0.05

statistical significance

→ low type 1 error probability

→ confidence that difference/effect exists

→ “not due to chance”

t-test

compares means of two groups

ANOV

compares means of 3 or more groups

hindsight bias

“I knew it all along” ← after conducting research and coming to a conclusion, you feel like this

confirmation bias

belief → seeks evidence that supports

belief → ignores evidence that goes against

construct validity

theoretical concept

does the test accurately assess the theoretical concept?

criterion validity

specific skills

does the test accurately measure a certain set of skills?

content validity

measure that accesses how well a test or other assessment tool covers all relevant aspects of a construct

do the test items collectively fully cover the theme being assessed?

concurrent validity

measure that accesses how well 2 tests or assessments agree when they were taken at the same time

does the test correlate well with an already established test?

predictive validity

ability of a test or measurement to predict a future outcome, such as behavior, performance, or disease

face validity

does the test look like a legitimate test?

test- retest reliability

does administering the SAME test give you similar results?

parallel forms reliability

do different test versions with similar content yield similar results?

inter-rater reliability

do obsevers of the same event rate it the same way?

split half reliability

is there consistency WITHIN the test?