final unit 256

mixed methods design

combining quantitative and qualitative methods

convergent parallel design

comparing and relating quan and qual then interpreting

explanatory design

Quan analysis followed by Qual (to explain Quan), then interpret

exploratory design

start with qual, use info to build to quan then interpret

qualitative designs

narrative research, phenomenology, grounded theory, ethnography, case study

phenomenology

shared experience between several people, interviews, common themes

narative research

1 or more people, life story of individual, many in depth interviews

grounded theory

in order to develop theory, stude a process/interaction, include many people, interview, how are common themes related

ethnography

culture based, immersive, study a culture itself, time intensive, live in the culture

Qualitative data

interview, observations, documents, audio-visual materials

p-hacking

re-analyzing data until find significant results(drop participants)

HARKing

Hypothesize-After-Results-are-Known

small samples are a problem because

they are more susceptible to skew

external validity comes from:

how sample is chosen, not size of sample

does theory testing mode have to generalize

nope, focuses on internal validity

frequency claims generalization mode?

always in this mode

association/causal claim generalization mode?

sometimes in this mode

types of replication

direct replication, conceptual replication, replication + extension

why study may not be replicable

contextually sensative effects, number of replication attempts lead for more chances of error, issue with original study

meta-analysis

what does the literature say, find average of all effects over all studies, a study of studies, strength/limitations, file-drawer issue

non equivalent control group post test only design

quasi experiment, no random assign, treatment/control

nonequivilent control group pretest post test

quasi, pre/post test, treatment/control. not random assign

interrupted time series design

quasi, DV measured repeatedly throughout the intervention, non manipulatable variable

nonequivalent control group interrupted time series design

quasi, treatment/control, dv measured repeatedly throughout the intervention, not random assign

internal validity threats in quasi experiments

selection effects, design confounds, maturation, history

quasi v. correlational

both independent groups, no random assign, no manipulated variables, Q: intentional selection of participants, causal argument, C: measuring to look at association

quasi experiment IV

participant variables, ex. age/gender, overall wants to find possible interventions

small n designs

very few participant, like low single digits

stable baseline design

small n, observe behavior at baseline then introduce intervention and observe

multiple baseline design

small n, stagger intervention across situation/time/contexts

reversal design

small n, have baseline/observe, introduce treatment/observe, take away treatment/observe

demand characteristics

when participants figure out what the study is about and change behavior accordingly, fix w/ double blind

weak manipulaiton null

doesnt cause enough change to bring desired effect

insensitive measure

pass/fail, low/med/hi, not sensitive enough to detect change

ceiling and floor effects

results cluster at top/bottom - too much w/in group variability

measurement error

not reliable/valid, fix w/ better measure, and measure more, bigger sample size

individual differences null

people spread out scores, it is hard to detect differences, fix w/ w/in groups design, more participants

situaiton noise

external distractors

what do you do with null results

you should still report them because it could help other people

w/in groups factorial

all groups experience all conditions

mixed factorial design

one V is independent groups, one is w/in groups, design is intermediate between w/in and independent groups for # of participants,

3 way interaction

are the 2 way interactions different in pattern? v1 influences how v2 and v3 interact

factorial in journals

___x____x____ design, significant main effects/interactions, significance, p-value, anova, f,

factorial in pop media

it depends, only when, to show interaction, participant variables - age, gender, ethnicity

factorial designs

study 2+ independent variables, to test limits of how IV affects different people groups, type of external validity, find moderators, can test theories

interaction

a difference in differences - effect of 1 - describe: one group was ___ than ___ group, especially true for ____

cell

unique condition, # of v1 levels x # of v2 levels

interaciton on a graph

if lines are parallel there is no interaction

main effect:

effect that individual variable has