research 2

nominal

no implied categories
no value
no order
ex- race, sex
non parametric

ordinal

adds meaning and order
rank
no equal distance
non parametric

interval

0 is arbitrary
equal distance but no 0
ex- temperature
parametric

ratio

0 means something
ex- weight and pulse rate
parametric

reliability

consistency

validity

accuracy

high validity means

high reliability

inter rater reliability

how consistent raters evaluate or judge a IV

test stability

whether the measure is stable over time
measuring something multiple times yields a consistent value
gets closer to the truth
stability over time

internal consistency

how well the items on the measure work together to produce similar scores
internal consistency

how to measure inter rater reliability

pearson product correlation or pearson's r

how to measure test- rest reliability

pearson product correlation or pearson's r

how to increase inter rater reliability

ensure proper training among raters
protocols are clear
rating is consistent

how to increase test stability

sufficient number of questions
ensure questions are easily understood
increase your sample size

face validity

does a test look like what it is supposed to test

content validity

addresses how well test questions match the content or subject area they are intended to asses
opinion based

predicative validity

tells you how well a certain measure can predicts future behaviors or performance

convergent validity

the degree to which a construct measure converges with other measures that should be measuring the same thing

discriminant validity

the degree to which a construct measure diverges from other measures that purport to be measuring something different

statistical significance testing doesn't tell us about what

strength or magnitude

random sampling error

the natural deviations that occur when randomly sampling from the population
expect fluctuation

bias

flawed sampling procedures whereby the researcher does not use a representative sample
not naturally occurring

sampling errors start with

random sampling error & bias

sampling distribution

the distribution of a sample statistic that would be obtained if all possible samples of the same size were drawn from a given population

central limit theorem suggests

the mean of the sampling distribution will equal the mean of the population
as the sample size increases, the standard deviation of the sample distribution decreases

statistical significance testing

quantify whether a result is likely due to chance (random sampling error) or due to some factor of interest

null hypothesis

no difference between the two groups

alternative hypothesis

there is a difference

type 1 error

reject the null when there is actually no difference between the two groups/ alpha/ level of significance
(rejected and shouldn't have)

type 2 error

when we fail to reject a null when it is actually false (doesn't reject the null when you should've)
beta

power

the ability of the test to correctly reject a null when it is false

a p value that is less than or equal to 0.05 or 0.01 is

statistically sig and you reject the null

a p value that is greater than 0.05 or 0.01 is

not statistically significant and fail to reject the null

as sample size increases ...

standard error decreases and distributions become more clustered around the population statistic

a really big sample can be

not clinically or practically useful

what is the big three

statistical significance testing
effect size
confidence intervals

Probability

communicated by statistical significance by telling us how likely the current result would be if the null were true

strength

communicated by effect size by telling us the magnitude of the experimental effect, or relationship, or odds between variables

precision

communicated by confidence intervals by providing a range of plausible values for the population parameter being estimated

effects sizes different tests

cohens d or hedges g
pearsons r or point biserial
odds ratio

cohens d or hedges g

standardized mean difference
estimate the magnitude of standardized differences between two group means

which has a bias correction for small samples

hedges g

cohens d or hedges g no effect

0.0

cohens d or hedges g small

+/- 0.2

cohens d or hedges g medium

+/- 0.5

cohens d or hedges g large

+/- 0.8

pearson's r

correlation/association
measures the strength of a linear relationship between continuous variables

point biserial

correlation/association
measure the strength of a relationship between one dichotomous and one continuous

pearsons r or point biserial no effect

0.0

pearsons r or point biserial small

+/- 0.1-0.3

pearsons r or point biserial medium

+/- 0.3-0.5

pearsons r or point biserial large

+/- >0.5

odds ratio

proportion
ratio of the probability that an outcome occurs to the probability that the outcome does not occur

odds ratio of 1.0

no difference

odds ratio greater than 1

positively related

odds ratio less than 1

negatively related

if the lower limit of the CI is greater than the effect size predicted by the null we can conclude that

the effect is real

CI establish

if an effect occurred
if effect was positive or negative
the precision of the estimate

standardized mean difference studies

null = 0

correlation/association studies

null = 0

proportion studies

null = 1

between subjects

subjects get 1 level of IV only, comparison is between different groups of same level

within subjects

all subjects are in all levels and are compared against themselves

if DV meets statistical assumptions the data is

parametrical (ratio interval)

qualitative research examines the

depth of a topic

quantitative research examines the

breadth of a topic

grounded theory

how a theory is generated and constructed

phenomenology

how someone makes sense and or gives meaning to things that happen

ethnography

describes aspects of a culture
researcher inserts themselves into the culture

types of basic qualitative research validities

confirmability, transferability, dependability, and credibility

confirmability

findings are shaped by subjects and not researchers bias, motivation, or interest

transferabilitiy

showing that findings have applicability in other contexts
can they describe the context of the study?

dependability

findings are consistent and could be replicated

credibility

confidential
truth of the findings

external validity best matches with

transferability

internal validity best matches with

confirmability

construct validity best matches with

dependability

audit trail

description of the decisions and steps researchers made through the project

peer debriefing

working with others who have an impartial view as a way to uncover bias and test the hypothesis

triangulation

multiple methods (sources or researchers) to confirm one common thing

prolonged engagement

cornerstone of qualitative research
spending time in the field to understand a culture, social setting, or interest

reflexivity

examining one's attitudes and judgments and how that affects the research process

theoretical assumptions

moveable target
getting the same theme of information from multiple sources- saturated
interview till no more new information

coding

trying to capture everything in an interview
looking for themes, topics
constantly compares

interviews emphasize

subjects POV

a telephone interview is usually what kind of data

quantitative

it is typical for a phone interview to be

structured

it is typical for a face to face interview to be

semi structured or unstructured

an interview should be what

more conversational

what questions should you avoid in an interview

leading

what kind of initial questions should you begin with

what events lead to
what was your life like

what are some intermediate questions

how did you feel about... when you first learned about it

what are good concluding questions

how have your views changes
what is advice you would give to someone

focus groups emphasize ...

a theme or topic and how people respond to each other

how many in a focus group

6-8 with a plus and minus margin of 2

tuskegee study

40 year study about syphilis where subjects did not know they had syphilis and were not given treatments when it became available
subjects were not allowed to quite

national research act 1974

1. established a commission to identify basic ethical principles
2. established the IRB

belmont report major 3

respect for persons
beneficence
justice

respect for persons

informed consent, privacy

beneficence

risk/benefit assessment
minimize risks