Research Methods Homan exam

this guide is bad, but usable ish i guess

Descriptive strategy

Only describes variables (well obviously but also) NO RELATIONSHIP EXAMINED BETWEEN VARIABLES

demand characteristics

seeing something or being in presence of something changes behavior (ex: hawthorne effect, weapons effect)

t test

(differences due to IV + differences due to chance)/differences due to chance

Time relatead threats for within subjects

History: environmental effects other than treatment that change over time

Maturation: changes in participants physiologically or psychologically

Instrumentation: u know

Regression towards mean: more extreme scores will be more average after being checked again

Order effects: Order of stuff can impact

Differential research

participants automatically in groups based on preexisting factors with goal to establish difference between them (between)

non experimental

Posttest-only nonequivalent control group design

uses preexisting groups (between), one group is treatment and one is similar but not equivalent for control. NO RANDOM ASSIGNMENT

XO

()O

quasi experimental

Pretest-Posttest nonequivalent control group

Compares two nonequivalent groups (experimental and control) NO RANDOM

Measures before and after, giving only one group treatment

OXO

O()O

quasi experimental

Pre-post designs

one group of participants, checks them twice (within)

Non-experimental

A nonexperimental pre-post design

One group (within), checks group before and after treatment administered

non exp because no attempt to control threats, just a small snapshot of a potentially much larger situation

Quasi-experimental pre-post design (Time series)

like nonexp prepost, in that it sees one group (within), but different because measures group multiple times both before and after treatment to check for trends

Cross-sectional

if it was not obvious, its between subjects of multiple age groups of people.

Problem is cohort effects: individuals who were born in same time period share beliefs that could be independent of just age

Longitudinal

within subjects, one group over many years, removes cohort effects however is much more costly,

problem is differential attrition: people leaving study, could skew results

Factorial designs

have more than one IV (factor)

think of box (2×2, 3×2, etc.)

Main effects

look at differences in levels of a single factor in factorial design, not examining relationship between the factors

Interaction

looks at relationship between factors to see of specific interaction causes something unique

remember, parallel lines mean NO INTERACTION, not parallel = INTERACTION

combined strategy (of factorial design)

factorial design where one factor is true and one is quasi

High order factorial design

3 factors (A B C) means 3 main effects, 3 two way interactions, and 1 3 way interaction

F ratio

(treatment effect + individual differences)/ individual differences

types of correlations

pearson: two continuous (scale) variables

spearman: two ordinal variables

point biserial: one continuous and one dichotomous

phi: two dichotomous variables

idiographic vs nomographic

ideo is individuals study, nomo is group study

IPA (Interpretive phenomenological analysis)

Goal is to explore how participants are making sense of personal and social world (NO TESTING HYPOTHESIS)

Reversal design

Single study, ABAB (A - baseline, B - treatment) demonstrates that treatment causes change

limits: not good with longer treatments, could be unethical to withdraw

Multiple baseline design

begin with two baselines phases, do treatments for each at different times

types:

• Subjects

• behaviors

• situations