psych midterm

Empiricism

the process of generating knowledge through the systematic use of one's senses

Theory

a set of statements that describe general principles about how variables relate to one another

Hypothesis

predictions about the outcomes of your research based on theory 

theory data cycle

Theory, research questions, research design, hypotheses, data 

Confound

alternate explanation for an outcome

research is

probabilistic, not deterministic

Cognitive Errors

Good story, availability heuristic, present/present bias

Motivated Reasoning

confirmation bias, cherry picking, bias blind spot

Validity

the degree to which a claim is accurate, well founded, or reasonable 

Construct Validity

how well is the variable measured 

External validity

do these results generalize

Statistical validity

are the results significantly significant

Internal Validity

does one variable cause a change in another

internal is only important for what claim

causal

(Within causal) Covariance

variables must be correlated with one another

(Within causal) Temporal precedence

causal variable must precede the outcome variable

(Within causal) Internal validity

no other variables can explain the relationship (3rd variable problem

Independent variable

manipulated

dependent variable

measured

control variable

constant

Internal-external seesaw

As we get more control (internal validity), we lose generalizability (external validity)

These are the three core ethical principles outlined by the Belmont Report

respect for person, justice, beneficence

Beneficence

do no harm; ensure research is beneficial

Justice

ensure some groups do not experience undue burden in research participation; ensure research benefits the group participating

Respect for Persons

informed consent, protection of special groups

Institutional review board (IRB)

Local committee charges with reviewing all human subjects research to protect the rights and welfare of participants

Fabrication  

making up data 

Falsification

manipulating pieces of the research process

Plagiarism

taking credit for work that isnt your own

Measurement

an Aspect of construct validity 

Conceptual variable

broad variable

Operational variable

specific variable

Nominal 

Levels are categories

Quantitative

Levels are numbers

Ordinal Scale

Categories with rankings

Interval scale

equal distance between observation

Ratio scale

equal distance and true zero

Reliability

consistency, preciseness, or dependability

Test retest

Looks at consistency over time

Interrater

Looks at consistency across 2+ coders/observers of behavior

Face validity (subjective)

Extent to which a scale appears to measure what is intended to measure

Content validity (subjective)

Extent to which a scale captures all parts of the construct its intended to measure

Criterion validity

Measures should correlate with behaviors/outcomes that are related to the construct

Known groups paradigm (a type of criterion validity)

Rather than using a behavior, are there existing groups that can help validate a measure

Convergent validity

Score on a new scale measuring depression are closely related to scores on an established depression scale

Discriminant validity

Scores on a new scale measuring depression are not related to scores on a scale of perceived physical health

Response set

a cognitive shortcut to avoid thinking about every question individually

Acquiescence

agreeing regardless of personal beliefs

fence sitting

staying neutral

social desirability

presenting in the favorable way

leading questions

questions we pre determined answers

barnum questions

vague statement that could apply to anyone

double barreled question

two questions in one

anchoring

relying to heavily on the first piece of information given

Observer bias

seeing whats expected

Observer effects

expectations → cues that affect behavior

Reactivity

subjects change behavior just due to being watched

simple random sampling

every individual from the population has an equal chance of being chose

cluster sampling

dividing a population into smaller, geographically or characteristically grouped clusters, and then randomly selecting a few of those clusters to study as a representative sample of the whole population

multistage sampling

taking participants from clusters as the groups get smaller and smaller

Stratified random sampling

researchers divide subjects into subgroups called strata based on characteristics that they share

Oversampling

incorporate more members of a certain community into your sample

Systematic sampling

a probability sampling method where researchers select participants from a population at regular intervals

Convenience sampling (biased)

a research method that involves selecting participants who are easy to access

Purposive sampling

a non-probability sampling technique where a researcher deliberately selects participants based on specific characteristics that are relevant to the study's objectives

snowball sampling

a non-probability sampling method that uses existing study participants to identify new participants

quota sampling

a non-probability sampling method that involves selecting a specific number of people from a population to participate in a study

Frequency claims

a statement that describes the rate or prevalence of a particular behavior, attitude, or characteristic within a population

Association claims

a statement that two variables are related to each other

causal claims

a statement asserting that one variable directly causes a change in another variable

Self report 

People tell you about themselves (lives, thoughts, behavior, opinions, etc)

Correlation coefficient

Statistic used to describe the relationship between two variables (r )

The farther away a correlation is from 0 the stronger it is ( r )

.1, .3, .5

Smaller CI/MOE

= more precise estimate

Can shrink CI/MOE by

increasing sample size

Online outlier

fall into the same pattern, but are more extreme (make the correlation seem stronger than it is)

Offline outliers

extreme and do not follow the pattern (make the correlation seem weaker than it is)

Multiple regression

mathematical way to examine the relationship between three or more variables 

Regression does not

establish causation

Statistical adjustment does not equal

temporal precedence 

Moderator

a variable on which the relationship between two other variable depends

Mediator

a way to understand “Causal chains” between variables

Bivariate

a study design which only 2 variables are measured

Multivariate

more than 2 variables are measures

Longitudinal

measured the same variable in the same group of people at different points in time 

Cross-sectional longitudinal design

2 variables measured over time

Autocorrelation longitudinal design

how a variable at 1 time point is related to another variable at another time point 

Cross lagged longitudnal story

measuring both variable at two different time and comparing the results to see which relationship is stronger 

Simple Experiments

At least 1 measured variable and at least 1 manipulated variable

Covariation simple experiments

Are differences in the DV related to differences in the IV

Temporal Precedence simple experiments

You need to manipulate the IV before the DV

Internal validity simple experiments

Ensure that the IV is responsible for the changes in the DV

Design confound threat to internal validity

variable that systematically varies with the IV

solution to design confound

think about confounds and plan in advance

Order effects threat to internal validity

behavior in one condition is biased by being in a different condition already

Practice Effects threat to internal validity

people might get better at the task or tired of the task (doing something gets easier over time)

Carryover Effects threat to internal validity

something about one condition changes your experience of the second condition (orange juice and toothpaste)

Solution for Order Effects

counterbalancing

Full Counterbalancing

–every possible order of the conditions

Partial Counterbalancing

only some possible order of conditions • Randomize presentation order