Exam 2- Lists and vs. Flashcards
Research designs that use an intervention? (3)
- experimental
- quasi- experi
- random control trial (RCT)
ways to control confounding variable?
- randomization
- crossover
- homogenity
- matching
- statistical control
3 types of causal relationships
- temporal
- relationship
- confounders
3 parts of every RCT
intervention, control group, participants randomly selected
C in PICO options (5)
current practice alternative treatment placebo attention control delayed treatment (*most ethical)
Limitations to RCT (3)
can’t test all interventions, cant randomize all variables, difficult admin approval
quasi experimental characteristics
- intervention
- no control group = comparison group
- no randomization
- offered to everyone in population
- less generalizability
- increase confounding variable
3 types of probability sampling +definition
- simple random-most basic probability sampling and involves using a tool to random select elements from a sampling frame
- stratified random- divides the population into two strata and then randomly selects samples from both strata.
- systematic- chooses to select a specific number such as every 4thperson to use in the sample.
- Probability sampling is the only viable method of obtaining a representative sample
2 types of participant responses to a question
- self report
- patient reported outcome
2 ways to determine effectiveness of tool
- reliability
- validity
characteristics of qualitative design
- emergent design evolves
- flexibility
- triangulation
- holistic
- ongoing data analysis
3 types of info ethnographers seek out
cultural:
behavior
artifacts
speech
3 ways to reduce bias
blinding
randomization
using a script
4 types of STUDY validity
internal
construct
external
statistical
ICES are Valid
characteristics of RCT
intervention
control group
randomized
primary
4 types of NONprobability sampling
consecutive-recruits everyone from an accessible population over a period of time
purposive-researchers handpick sample members based on traits in the population
convenience-most conveniently available people as participants
quota-strata information to figure out how to use a balanced sample
PCCQ
prob vs non prob sampling
PROB
- simple random
- stratified random
- systematic
NON prob
- consecutive
- purposive
- convenience
- quota
4 steps of descriptive phenomenology + definitions
- bracketing-process of identifying and holding in abeyance preconceived beliefs and opinions about the phenomenon understudy (reflexive journaling)
- intuiting-remaining open to the meanings attributed to a phenomenon by those who have experienced it
- analysis-extracting significant statements, categorizing, and making sense of essential meanings
- describing-understanding and defining the phenomenon
interpretive vs descriptive phenomenology (the types of phenom studies)
interp: AID, hermeneutics, understanding human experience, Van Manen, Heidigger, Benner
descrip: BAID, essence, and Husserl, Colaizzi, Van Kaam, Giorgi
2 types of phenomonlogical studies?
interpretive, descriptive
3 resons to used mixed method?
- complimentary
- practicality
- enhanced validity
“When I feel mixed up, I practice complimenting myself enhance my validity “
5 types of mixed studies- how are they written?
concurrent: QUAN/QUAL, QUAN/qual, QUAL/quan
sequential: QUAN -> QUAL, QUAL + quan
convergent: QUAL + QUAN
explanatory: QUAN -> qual or quan -> QUAL
exploratory: QUAL -> quan or qual -> QUAN
4 phases of drug trial
1: develop
2: pilot (small)
3: RCT
4: gen pop
quantitative study bias types:
contamination of treatment, noncompliance bias, and attrition bias, sampling (volunteer/nonresponsive), measurement (acquiescence, observer)
4 methods of spradley
domain
taxonomic
componential
theme
“The theme of Spradleys domain is taxonomic components”
3 parts of data triangulation
time, space, person
no Time for teePee in Space with your person
cross-sectional vs longitudinal data collection
cross sectional = one specific time
longitudinal =across extended period of time or multiple times
prospective vs retrospective study?
retro = looking back prospect = cohort , looking forward, begins with presumed cause and looks for presumed effect
descriptive study vs descriptive phenomenology
descriptive = study observes the prevalence of something without determining a relationship or linking a cause
-If a article identifies itself as a qualitative study but does not fit a particular design, it is
called a descriptive qualitative study.
descriptive phenom = describing the nature of the phenom, BAID, essence, and Husserl, Colaizzi, Van Kaam, Giorgi
types of non-experimental studies (4) + definitions?
- correlation=determines an association but does not prove cause and effect
- case-control=begins with the effect or outcome, look retrospectively at patient data to observe for potential causative variables
- prospective= begins with a presumed cause, looks for a presumed effect.
-descriptive= observes for the prevalence of something without determining a
relationship
interview vs questionnaire
interview: self-reported interview data is collected with a formal written document verbally in person or by telephone
questionnaire: participants complete a form or survey themselves.
visual analog vs likert scale
visual analog: measures subjective experiences such as pain or fatigue from 0-10
likert: scale that has a rating with options such as strongly agree, agree, neutral, disagree, strongly disagree, or always, sometimes, never, n/a
self reports have what 2 kind of errors?
accuracy
validity
2 types of biophysiologic measures?
in vivo : performed directly on the participant to obtain a value, such as a blood pressure.
in vitro: takes something from the participant and analyzes it, such as drawing blood to measure lab values.
reliability vs validity
Reliability: extent to which scores are free from measurement error
validity: degree to which an instrument measuring a concept is appropriate to measure
3 types of MEASUREMENT validity
face: an instrument looks like it measures the target construct
content: extent an instrument’s content adequately captures the construct
criterion: the extent to which scores are a good reflection of an ideal measure of the construct
The CONTENT of my FACE makes me want to CRI
sample size grounded theory vs phenomenology
grounded = 20-30 ppl phenom = <10
photovoice vs photoelicitiation
Photoelicitation: uses photos to conduct an interview
Photovoice: involves the participants themselves taking pictures and interpreting them.
emic vs etic
emic= insider
eTic- ouTsider
nominal vs interval vs ordinal
- nominal : statistical analysis of quantitative data, this measurement is the lowest level and involves using numbers to designate attributes (ex: brown eyes 1, blue eyes 2, green eyes 3)
- interval: statistical analysis of quantitative data, this measurement ranks people on attributes w/a # that specifies a distance between the 2 (ex: IQ test)
- ordinal: statistical analysis of quantitative data, this measurement ranks ppl based on an attribute (ex: ADLs: 4 = independent, 1 = completely dependent)
crossover vs matching vs homogenity
crossover: giving more than one treatment, but separate them out and use randomization to determine ordering
matching: developing a comparison group with the same confounding variables as the control group
homogenity: a method of restricting a study to individuals who have confounding variables
statistical heterogeneity vs effect size - meta analysis or meta synthesis?
-statistical heterogeneity: factor assessed to determine if a meta-analysis is APPROPRIATE
(are the results conflicting between different studies on the same research question?)
if they have conflicting results, it’s NOT appropriate to use a meta-analysis
-effect size: meta-analysis, this is used to determine the RELATIONSHIP between variables across multiple studies
types of mixed method research
- outcomes:focuses on examining the effect of healthcare on the outcome of patients
- health services: this type of research examines how organizational structures and process, health technologies, social factors, and personal behaviors affect access to health care, the cost and quality of health care, and people’s health and well-being
- evaluations: does not seek to increase the body of research knowledge, but affirm or approve a potential quality improvement initiative
- nursing intervention:involves the use of complex interventions (rather than 1 like a clinical trial)
- survey research:research obtains quantitative information about variables in a population (usually non-clinical)
clinical vs statistical significance
- clinical: practical importance of research results in terms of whether they have genuine palpable effects on the daily lives of patients or on the health care decisions made on their behalf
- statistical: the results are not likely to be due to chance.
ways to measure clinical significance (group vs individual)
change scores +
- GROUP LEVEL: effect size indexes, confidence intervals, and number needed to treat
- no p score
-INDIVIDUAL level: benchmark, minimal important change
3 ways to determine minimal important change
(individual clinical significance)
- benchmark
- ask patient what they think
- count 0.5 SD in a measure as clinically significant
what is a type 2 error? how do we reduce risk of having it?
assumes the null hypothesis is true when in fact the independent variable did have an effect
(false negative)
-increase sample size = reduce risk of type 2 error
2 ways we know something is statistically significant?
- values that fall within 2 standard deviations of the mean.
- Confidence intervals of 95% or greater
3 tests for statistical significance
- T test (indep. or paired) - testing differences in two group means- gives us t score
- ANOVA-test mean group differences of three or more groups – gives us F value
- Chi-squared- focuses on the difference in proportion
where do we get an R value? range?
multiple correlation coefficient
range: 0-1
^ R = more likely the predictors accounted for the variation
3 types of multivariate statistical analysis
- Multiple regression = R value
- ANCOVA = ANOVA + mult regression
- Logistics regression =yield an odds ratio called a OR for each independent variable as well as a confidence interval
3 tests for reliability
- Test-Retest Reliability= Intraclass correlation coefficient (ICC)
- Cohen’s kappa = Interrater reliability = would 2 people rate thing same way
- Coefficient alpha or Cronbach’s alpha =• Internal consistency reliability= measure how often components of a multicomponent tool measure the same attribute.
For all of these, scores closer to 1.0 means a higher level of reliability.
2 terms that fall under criterion validity
sensitivity and specificity
