experimental psych exam 1

commonsense psychology

nonscientific data gathering

• uses nonscientific sources of data and nonscientific inference

scientific mentality

behavior follows a natural order and can be predicted

empirical data

observed, experiences, tested “research”

• scientific empirical method → true research

law

consists of statements generally expressed as equations with few variables that have overwhelming empirical support

• psychology doesn’t have laws

theory

a broad set of ideas that are related to one another supported by a series of tests

• i.e.: cognitive dissonance theory, bystander effect, halo effect

hypothesis

a specific proposition taken from a theory

• the “driver” of the study / the “center point” of the paper

• everything centers around the hypothesis

• you should be able to tell the IV, DV, and direction of the effect

good thinking

critical to the scientific methos

• engaged when data collection and interpretation are systematic, objective, and rational

systematic

doing things the same way for everybody as much as possible

objective

no biases/wants

• for the purpose of finding the answer of a question

rational

not overstepping the limits of what the study is telling you

weight of evidence

how science advances

• keep studying until one side weighs heavier than the other

replication

an exact or systematic reputation of a study

direct replication

• not chosen very often

• harder to get published

conceptual replication

• same concepts but different operational definitions

• the same variables adds to generalizability

replication-plus-extension

• direct or conceptual extension

• adds a level to the variable

• most willing to publish

prediction

knowing in advance when behaviors should occur

control

experimentation

• the use of scientific knowledge to influence behavior (manipulating variables)

applied research

solving specific, real-world problems

• i.e.: health market research → the largest industry to partake in applied research

basic research

aimed at theory development

• i.e.: attractive female faces attract everyone

measurement

operational definition (numeric)

• i.e.: attachment style

• “how do i make this into a number?”

correlation ≠ causation but …

experiments can establish causation

what does it mean when you have a confound

no causality

what can stop a experiments/hypotheses from being tested

ethical limitations, financial/resource issues, feasibility, not able to test what you want to

what are extraneous variables

confounds

quasi means…?

seeming like

• something that might form groups

• not an experiment because there’s no random assignment

quasi variable

treat like a manipulated variable but you can’t really manipulate it

• i.e.: participant variables (gender, hair color, height) are quasi-variables

preexisting antecedent conditions

life events or subject characteristics on behavior

• there’s no manipulation so it = x

• manipulation causes causality

when should we use quasi experiments

when we can’t or shouldn’t manipulate antecedent conditions

• quasi experiments could study the effects

pearson’s r

from -1 to +1

• perfectly correlated = perfectly straight line — not entirely possible

• correlations coefficients of 0 = no relationship — visually seen as scattered data

strength

visually based on how closely data resembles a straight line

• how far away from 0 — NOT THE SAME AS SLOPE

+0.63 or -0.72 — which is a stronger correlation?

A: -0.72

• why?: because it’s the closest to zero

range truncation

a restriction of range

• misleads you

• very bad

• the research question dictates whether or not to include extremes or not

coefficient of determination (r² )

how much variability in the DV explains the IV

what are the three issues of (r² )?

1. causal direction

2. bidirectional causation

3. 3rd variable problem

causal direction

since correlations are symmetrical, A could cause B just as readily as B could cause A

• i.e.: does insomnia cause depression or does depression cause insomnia?

bidirectional causation

two variables may affect each other

• i.e.: insomnia and depression

3rd variable problem

a third variable may create the appearance that insomnia and depression are related to each other

• i.e.: family conflict to insomnia and depression

• this is different from a confound

confound

when manipulating the IV, accidentally manipulating other variables and cant find what the main DV is

multiple regression

predicting something and using at least 2 predictors to understand their combined effect and individual contributions

• we can only control for what we measure

• doesn’t solve problem for correlational data

causal modeling

• done with correlational data

• when we can’t do an experiment

• can’t establish causation

• suggests cause-and-effect relationships between behavior

what are two forms of causal modeling

1. path analysis

2. cross-lagged panel designs

path analysis

• measures every variable at once

• every variable that is given an arrow is a predictor

• every variable that is receiving an arrow is a DV

cross-lagged design

measures relationships over time (longitudinal), with usually 2 variables at a time

• can’t do an experiment

experimental hypothesis

a manipulated hypothesis

nonexperimental hypothesis

ask: is it manipulated or not?

testability

without testability, we can’t evaluate the validity of a hypothesis

parsimony

we prefer a simple hypothesis over one big, confusing one

• a simple hypothesis allows us to focus our attention on the main factors that influence our DV

induction

process of reasoning from specific cases

• theory based

deduction

top-down processing; general principles and use logic to reach specific conclusions

solomon 4 design

1. a group that received the pretest, treatment, and posttest

2. a nonequivalent control group that received only the pretest and posttest

3. a group that received the treatment and posttest

4. a group that only received the posttest

longitudinal studies

the same group of subjects is measured at different points of time to determine the effect of time on behavior

cross-sectional studies

subjects at different developmental stages are compared at the same point in time

pretest/posttest design

a researcher measures behavior before and after an event

• this is quasi experimental

• DV → take your manipulation (IV) → then DV again

practice effects

performance improvements resulting from repeating tasks or tests