EXAM 2 Flashcards
population
group that meets study criteria
elements
basic unit of the population (what is being studied?)
subjects
select group of subjects who will represent all eligible subjects
target population
elements that meet all study criteria
accessible population
group of elements that the researcher has access to
sample
select group of subjects who will represent all eligible subjects
sampling error
subjects don’t represent population
(often a result of small sample size)
sampling bias
sample over or under represents characteristics of target population
inclusion criteria
those things people have that you want in your study
exclusion criteria
those aspects to people you do not want in your study
power analysis
analysis that shows how large a sample needs to be in order to detect a difference in the outcome
validity
accuracy
does the instrument accurately measure what it is supposed to measure
reliability
consistency
extent to which the instrument produces the same results if behavior is repeatedly measured with the same instrument
stability
same scores with repeated tests
equivalence
agreement between raters or similar test produce same results
homogeneity/internal consistency
all items in a questionnare measure the same concepts
external validity
the degree to which the results of the study are generalizable to other populations
internal validity
did the independent variable really make an impact or were there confounding factors
content validity
does the instrument adequately represent the content
CVI
content validity index (CVI)
tells us: how accurately the question asks what we want (0.78-1.0 is acceptable)
criterion related validity
how much do the observed score and the true score relate to eachother
construct validity
how well does the instrument measure a theoretical concept
cronbach alpha
probability of making a type 1 error (tests internal reliability)
used for the likert scale
minimum acceptable 0.7
probability sampling
every element has an equal chance of being selected, done w/ randomization
simple randomization
most effective yet time consuming (draw names out of a hat)
stratified randomization
divide population into strata, subjects were randomly selected from each strata
cluster sampling
multistage sampling
(ex: BSN students in US randomly choose 10 states, then 3 schools from each state, then % of students from each school)
nonprobability sampling
random selection not required
less likely to be representative of whole population
convenience sampling
“accidental”, easy access inclusion criteria determined before selecting subjects
quota sampling
elements are conveniently chosen
recruitment until target sample size is reached
purposive sampling
subjects selected who are considered to be typical of the population
useful in studying populations w/ rare characteristics
snowball sampling
(networking) identify initial participant who then refers researcher to others who meet criteria for the study
relationship btw reliability and validity
CAN be reliable without validity
CANNOT be valid without reliability
descriptive statistics
allows researchers to describe and summarize data
measures of central tendency
mean- arithmetic average
median- score with 50% above and 50% below
mode- most frequent value in a distribution
measures of variability
range- difference between highest and lowest scores
standard deviation-measure of average deviation of the scores from the mean and should always be reported with the mean
inferential statistics
looks at error
allows the testing of hypotheses using data obtained from samples
parameter vs statistic
parameter- characteristic of a population
statistic- characteristic of a a sample
measurement
process of assigning numbers to variables or events according to rules
nominal
lowest level of measurement
classifies variables into categories (dichotomous- yes/no, female/male) (ex: apples, oranges, lemons)
ordinal
relative rankings of variables (1st, 2nd, 3rd)
interval
scale with equal intervals and NO absolute zero
can be positive or negative (temp)
ratio
highest level of measurement
scale of equal intervals and absolute zero
must be positive (height, weight, BP, pulse)
hypotheses
statements about the researcher’s prediction of the relationship between variables in a specific population
research/scientific hypothesis
what the researcher believes will be the outcome of the study
null hypothesis
says the relationship does NOT exist
researcher either accepts or rejects the null
type 1 error
researcher rejects the null when the null is true
the worst, gives false hope
type 2 error
researcher accepts a null when it is actually false
a missed opportunity
statistical significance
unlikely to have been caused by chance
P value: tells us probably of error occurring
parametric tests of significance
answers whether null is to be accepted or rejected
most powerful, gives effect of intervention
used with interval and ratio variable
variable must be normally distributed
pearson R
Pearson R
-1 to +1
(closer to 0=weaker the relationship)
for parametric testing
nonparametric tests of significance
for nominal and ordinal data
used when data is skewed
assesses relationship, not effect
tests of difference
T-test: statistically tests mean difference between 2 groups. only used with parametric testing
Degrees of freedom: tells us variation within a sample (n-1, total # of variables-1)
fisher vs chi square test
fisher- smaller samples, less than 6 in each cell
chi square- look at difference in frequency btw big groups
confidence interval (CI)
if it crosses 0 or 1= NO significance
also,95 or 99%+
tests of relationship
explore association or correlation between two or more variables
systematic reviews
pulling together a collection of studies
meta analysis
critical appraisal w/ statistical analysis
level 1
goal: determine effect of IV on DV
forest plot!
phase 1 of meta analysis
extract data
phase 2 of meta analysis
decision about appropriateness of calculating pooled average
effect size
estimate of how large a difference there is btw intervention and control groups
meta synthesis
integration of qualitative studies
generalizability
ability to apply results of study to other similar populations
epidemiology
study of the distribution of disease
prevalence
cases that exist in the population
incidence
new cases
risk ratio
cohort studies
follows group over time to see who develops outcome (disease)
tells us the risk of getting disease when exposed compared to those not exposed
risk ratio calculation
(subjects with exposure and disease/total exposed) over / (subjects with no exposure and disease/total not exposed)
A/(A+B)over C/(C+D)
RR=1 (no association)
RR less than 1 (lower incidence in exposed group)
RR=1+ (possible risk factor)
odds ratio
case control studies
tells us the risk of having outcome when exposed
odds ratio calculation
(disease and exposed/disease and not exposed)= X
(no disease and exposed/no disease and not exposed)= Y
X/Y= odds ratio
A/C=X B/D=Y
answer is “blank times more likely that they were exposed and developed the disease”
quality improvement
uses data to monitor the outcomes of care processes and improvement methods to design and test changes to continuously improve the quality and safety of health care system
CONTINUOUS IMPROVEMENT
benchmarks
goals that are set to determine at what level the outcome indicators should be met
PICO
Population
Intervention
Comparison
Outcome
pyramid of evidence
1- meta analysis
2- randomized control trials
3- quasi experimental
4- nonexperimental
5- metasynthesis
6- qualitative
7- opinions by committees or organizations
process of EBP
Ask
Acquire
Appraise
Apply
Assess
informed consent
requires…
Information
Comprehension
Voluntariness
qualitative study
words, feelings, descriptions
goal is to understand
(metasynthesis)
quantitative study
numbers
tests an intervention
autonomy
respect for persons
self determination
beneficence
to do good
justice
fair treatment
nonmaleficence
do no harm
fidelity
truthfulness
