Unit 3: Ch. 10 Flashcards
population
the aggregate of cases in which a researcher is interested
“aggregate” = everybody in the population
sample
the selection of the portion of the population
-subgroup of the population
"N" = entire sample "n" = subgroup of entire sample
eligiblity criteria
characteristics that define the population
Types:
- inclusion criteria
- exclusion criteria
inclusion criteria
characteristics that you want your sample to have for being in the study (what qualifies them to be in the study)
ex: postpartum study - you want to study only first time adolescent moms ages 14-19
exclusion criteria
characteristics that your sample may have that keeps them out of the study (things that disqualify them/exclude them from being in the study)
ex: pregnancy complications such as preeclampsia - if any of the moms in the study had preeclampsia, that excludes them from being in the study
strata
subpopulation of the population (“layers”)
ex: dividing one group further into subgroups
sometimes males and females in each group
target population
entire population of interest
ex: every adolescent mother in the US
accessible population
proportion of the target population who is actually accessible to the researcher
ex: target population is every adolescent in the US but the researcher can only access the ones within 40 miles
representative sample
a sample whose key characteristics closely approximate the population
-want your sample to be as representative of the whole population as possible
ex: our class is representative of BSN students general but our class is NOT representative of the whole university in general (b/c there are various majors at MU)
representative sampling is most easily achieved with…? (3)
- probability sampling
- homogenous sample/population
- larger samples
probability sample
random selection or random assignment
homogenous sample/population
where population is alike on key characteristics
larger samples
more of whatever the characteristic is that you’re looking for
-but you don’t want your sample to be too large (largeness has its limits)
sampling bias
the systematic over- or underrepresentation of segments of the population on key variables when the sample is NOT representative
happens when the sample is NOT representative
sampling bias slants your results one way or the other so that the results aren’t actually what you think they are
what are 2 sampling designs in quantitative studies?
- nonprobability sampling
2. probability sampling
nonprobability sampling
involves nonrandom selection; 5 types exist
- Convenience sampling: people who meet the inclusion and exclusion criteria but they’re people who are convenient for the researcher to access
- ex: collecting data at a clinic - the convenience sample would be whoever shows up at the clinic that day
- The most common type of sampling used in nursing; most vulnerable to bias. - Snowball sampling: people in the study help you recruit others to be in your study
- used more in qualitative studies - Quota sampling: recruitment of a certain subgroup based on certain characteristics to represent the sample (may have a number in mind of how many people you need)
- keep sampling until you reach your quota - Consecutive sampling: recruitment of people from the accessible population who meet the eligibility criteria for a certain period of time
- Purposive sampling: recruitment of participants based on personal judgment of the researcher about who will be the most informative; used a lot in qualitative studies
probability sampling
involves random selection of subjects and each subject has an equal, independent chance of being selected; 4 types exist
- Simple Random sampling: selection of sample via a random procedure
- ex: pulling names out of a hat - Stratified Random sampling: subdivisions of the population according to some characteristic
- ex: race, gender - Cluster sampling: multi-stage sampling in which large clusters/groups are selected with successive subsampling of smaller groups
- ex: going to do a study. Randomly selects 5 states out of 50 that you want to get your sample from. In each of those 5 states, randomly select 1 city to collect sample. Within each city, select a single hospital; within each hospital, select a single unit to actually collect the data from - Systematic sampling: where every 5th, 10th, nth person is selected (researcher decides how many)
Snowball and purposive sampling are mostly used in ____ studies.
qualitative
What are the 5 types of nonprobability sampling?
- convenience sampling
- snowball sampling
- quota sampling
- consecutive sampling
- purposive sampling
What are the 4 types of probability sampling?
- simple random sample
- stratified random sampling
- cluster sampling
- systematic sampling
sample size
defined as the number of study participants in the final sample
- sample size adequacy: need a sample that’s the perfect size
- sample size needs can and should be estimated through a power analysis
power analysis
purpose of a power analysis is to determine the appropriate sample size
sample size too small: sample may not have characteristics you need
sample size too big: the characteristics you’re looking for will eventually emerge b/c the sample is too big
The basic decision in the types of data collection is the use of?
- new data
2. existing data
new data
collected specifically for research purposes (primary data, primary analysis)
existing data
records (e.g. pt charts)
historical data
existing data set (secondary analysis)
it’s ok to use existing data in qualitative studies
what are the 3 types of data collection methods?
- self reports
- observation
- biophysiologic measures
self-reports
Participants themselves report on whatever it is. These data are collected with formal instruments such as questionnaires or interviews. There are some things that only the participant can tell you what it is.
observation
researcher watches or observes events under investigation
biophysiologic measures
bodily functions (BP, pulse, temp)
what are the 2 types of self-report measures?
- interviews
2. questionnaires
interviews
interview questions/schedules are prespecified
- asked orally
- advantages: higher response rate; may be more appropriate for diverse audiences; time in interviews to clarify questions; more opportunity to get supplementary data through observation
questionnaires
questions that are prespecified on a written form and self administered. Can be closed ended or open ended questions
- closed ended questions (fixed alternatives): looking for a specific answer
- open ended questions: much more general
- advantages: cheaper than interviews; less time consuming than interviews; better possibility for anonymity and greater privacy
questionnaire scales
used to make fine quantitative discriminations among people with different attitudes, perceptions, traits (many types)
- ex: how anxious are you on a scale of 1-7?
- likert scales
- semantic differential scales
- visual analog scales
likert scale or likert-type scale
usually 5 or 7 responses for each item
scores assigned to each response
each individual question may be referred to as an item
responses vary (ex: never, sometimes, always…)
depends on the researcher
semanitc differential scale
continuum between 2 adjectives (2 extremes)
continuum varies from 5 to 9
visual analog scale (VAS)
pictures and numbers are used to measure the variable
the number continuum varies
visual scales are good for people who don’t speak English
response set biases
tendency of some people to respond to some items in characteristic ways
-respond to characteristic ways independent to content (doesn’t matter what they’re asked - they answer in certain ways)
examples:
- social desirability response set bias
- extreme response set
- acquiescence response set (yea-sayers)
all of these slant your results; you’re not getting the truth of how the participants feel
data can come out completely different (from what you expected) if the people answer like this
-may have to throw out a person’s results if they always answer like this
social desirability response set bias
person tends to answer in politically/socially correct ways
extreme response set
people who answer that everything is really good or really bad (don’t give a lot of in-between responses)
acquiescence response set (yea-sayers)
the people go along with what the majority says they should go along with (whatever they think the majority would say)
evaluation of self reports
want them to be strong on directness
it allows access to info otherwise not available to researchers (ex: questionnaires)
can’t always be sure subjects feel the way they say they feel
may be sure of the reliability and validity of the tool you’re using to measure results
-does it measure what it purports to measure? how consistent are the results?
what are the 4 types of observations?
- structured
- rating scales
- time sampling
- event sampling
structured observations
something such as a checklist; formal systems for recording incidents; specify behavior/event
only check the things you see
checklist means you’ll be looking for a lot of things
rating scales observations
ratings are on a descriptive continuum, typically bipolar
may be looking at severity of behavior
an event happens and you rate how severe it is (ex: one kid hits another kid and you have criteria for determining the severity)
time sampling observations
looking at how often a particular behavior is happening; how many times a behavior happens w/in a specific length of time
event sampling observations
looking for important/specific events
-ex: may be looking for who is bullying who, how many kids in a playground bully…
biophysiologic measures (2 types)
- In vivo measurements: nothing ever leaves the body
- may be something like BP or temp measurements - In vitro measurements: things performed outside the person’s body; something leaving the person’s body
- ex: taking blood sample, urine sample; spit
Both tend to be strong on validity, objectivity, and precision
Not always suitable for every study or QI project
sampling plan
specifies in advance how participants will be selected and how many to include
elements
basic units
in nursing research elements of research are usually humans
____ sampling is the only viable method of obtaining representative samples
probability
____ samples are rarely representative of the entire population
nonprobability
sampling error
refers to the differences between population values and the sample values
e.g. the average age of the population and the average age of the sample
sampling interval
standard distance between selected cases
used in systematic sampling (a type of probability sampling)
critiquing sampling plans (5)
- type of sampling approach used (e.g. convenience, consecutive, random)
- population under study and eligibility criteria for sample selection
- sample size w/ a rationale
- description of sample’s main characteristics (e.g. age, gender, clinical status, etc.)
- number and characteristics of potential subjects who declined to participate
response rates
number of people participating in the study relative to the number of people sampled
nonresponse bias
differences between participants and those who declined to participate
aka response bias
scale
device that assigns a numeric score to people along a continuum
like a scale for measuring weight
summated rating scales or composite scales
a person’s total score is the sum of item scores and these scales are sometimes called summated rating scales or composite scales
in observational studies, researchers have flexibility with regard to several important dimensions (3)
- focus of the observation
- focus can be broadly defined events or on small, highly specific behaviors - concealment
- researchers don’t always have to tell people they’re being observed - method of recording observations
- observations can be made through human senses, recorded on paper, or done with equipment (ex: video equipment, computers)
reactivity
behavioral distortions due to the known presence of an observer
category system
represents a method of recording in a systematic fashion the behaviors and events of interest that transpire w/in a setting
checklist
the instrument observers use to record observations
rating scale
an instrument that requires observers to rate phenomena along a descriptive continuum
can be used as an extension of checklists
time sampling
involves the selection of time periods during which observations will occur
time frames may be selected systematically (e.g., every 30 seconds at 2-minute intervals) or at random
event sampling
researchers select integral behaviors or events to observe
requires researchers to either have knowledge about the occurrence of events (e.g., nursing shift changes) or be in a position to wait for their occurrence
factors that interfere with objective observations (3)
- emotions, prejudices, and values of the observer may lead to faulty interference
- personal views may color what observers see in the direction of what they want us to see
- anticipation of what is to be observed may affect what is perceived
closed-ended questions
used in self-reports; aka fixed alternative questions
ones in which the response alternatives are prespecified
open-ended questions
used in self-reports
allow participants to respond to questions in their own words
advantages of biophysiologic measures (4)
- relatively accurate and precise, especially compared to psychological measures, such as self-report measures of anxiety or pain
- objective - ex: 2 nurses reading from the same spirometer output are likely to record identical tidal volume measurements, and two spirometers are likely to produce the same readouts
- pts cannot easily distort measurements of biophysiologic functioning
- provides valid measures of targeted variables: thermometers can be relied on to measure temperature and not blood volume, and so forth. For nonbiophysiologic measures, there are typically concerns about whether an instrument is really measuring the target concept