Midterm Flashcards
What is nominal data?
categorical
-show differences in populations
ex. hair color or gender
Statistical tests for nominal data
independent- chi square
dependent
- McNemar or Cochran’sQ
Null hypothesis of nominal data
X2=0
Hypothesis of nominal data
X2 is not equal to 0
What type of data is ordinal data?
numerical scores to make a ranking over set of data points
ex. severity of disease or pain scale
Statistical tests for ordinal data sets
independent- mann-whitney or kuskal-wallis
dependent- wilcoxin matched pairs or friedmans
chi-square
Null hypothesis of ordinal data
mean difference = 0
hypothesis of ordinal data
mean difference is not equal to 0
What is continuous data?
numerical scores that are infinite and have meaning; shows difference in means
ex. pocket depth and temperature
Statistical test for continuous data:
independent- T test or ANOVA
dependent- T test or repeated measures ANOVA
Null hypothesis for continuous data
mean difference = 0
hypothesis for continuous data
mean difference is not equal to 0
What is risk ratio or relative risk?
the probability of an event occurring, shows the strength of association between exposure and disease
-used in RCT and cohort
What does a relative risk of 1 mean?
there were not differences between groups
What does a relative risk less than 1 mean?
risk of bad outcome is decreased
What does a relative risk greater than 1 mean?
risk of bad income is increased by the intervention
What is a Odds Ratio?
estimates risk ration
-comparing the odds of an event in one group to the odds of an event in a comparison group
-commonly used in case control studies in epidemiological research
What is absolute risk reduction?
difference in outcome rates between the control and experimental groups
-change in risk of a given intervention in relation to the control
-inverse od NNT
What is number needed to treat?
the number of patients that need to be treated in order for one patient to have a benefit
What is the ideal number needed to treat?
1
What does higher number to treat mean?
treatment is less effective
What does a negative number needed to treat mean?
intervention is harmful
What do confidence intervals give?
provide information about the most likely range of population parameter
-actual range is the odds ratio
What is the variability of confidence interval called?
standard deviation
Small sample size usually generates…
wider range of intervals
What is sensitivity?
probability that a subject with the disease will screen positive
What is specificity?
probability that a subject who is disease free will screen negative
How to calculate sensitivty:
true positive/ true positive + false negative
How to calculate specificity?
true negative/ true negative + false positive
How to calculate PV+
true positive/ (true positive + false positive)
How to calculate PV-
true negative/ (true negatives + false negatives)
What should we do to prevent cognitive bias?
-always question assumptions
-keep an open mind
-always question the evidence
What is logical fallacy?
error in logical argumentation
What is cognitive bias?
genuine deficiency or limitation in our thinking
-a flaw in judgement that arises from errors of memory, social attribution, and miscalculations
What is the bandwagon effect?
tendency to do (or believe) things because many people do (or believe) the same
What is confirmation bias?
looking for and valuing information that confirms what we believe
-misleads us about what is true
What is gamblers fallacy?
putting tremendous amount of weight on previous events
What is interviewer bias?
the distortion of response to an interview which results form differential reactions to the style and personality of the interview or to the their presentation questions
What is normalcy bias?
refusal to plan for, or reaction to, a disaster which which has never happened before
what is measurement bias (systemic error)?
systemically overstating or understanding the true value of a measurement
What is negativity bias?
paying more attention to negative news or perceiving it to be more important
What is observation selection bias?
suddenly noticing things that you did not notice before and wrongly assuming the frequency has increased
What is performance bias?
when one group of subjects gets more attention than another group which results in differences between groups
What is positive expectation bias?
a sense that our luck has to eventually change
What is publication bias?
when the outcome of a study influences the decision whether to publish it
What is question order?
inadvertently influencing responses due to the order questions are asked
What is recall bias?
participants do not remember previous events or experiences accurately
What is respondent fatigue bias?
phenomenon that occurs when survey participants become tired which results in the quality of data deteriorating
What is response bias (survey bias) ?
tendency for participants to answer questions untruthfully or inaccurately
What is sampling/selection bias?
sample obtained is not reflective of the target population
What is sponsor bias (funding bias)?
Study outcomes that support the interests of the study’s financial sponsor
What is status quo bias?
making choices that guarantee things remain the same or change as little as possible
What is verification bias?
outcomes more likely to be found in treatment group due to investigators knowing which person is in experimental and control groups
What are the clinical study categories?
etiology, diagnosis, therapy, prognosis, and clinical predication guides
What does the P stand for?
people
What does I stand for?
intervention
What does C stand for?
comparison
What does O stand for?
outcome
What are two most important terms in PICO?
intervention and comparison
What is quantitative data?
numbers
What is qualitative data?
can be interpreted
What is experimental research?
-quantitative
-investigate cause and effect
-investigator controls variables
-randomized control trial
What is observational research?
-quantitative or qualitative/ quasi
-prospective vs retrospective
What is an independent variable?
changes/ investigator controls
What is a dependent variable?
investigator does not control
What are confounded variables?
error
-not measured by researcher and has impact
What things affect validity?
internal- selection, maturation, and instrumentation
external- publication and financial
What does null hypothesis mean?
reject hypothesis
What is a type I error?
false positive
What is a type II error?
false negative
What is necessary to calculate the effect size?
standard deviation and mean to calculate
What is evidence based dentistry?
use current scientific evidence to guide decision making in dentistry
3 part harmony
-best evidence
-clinical judgement
-patient values
order of heirarchy of evidence:
-meta analysis
-systemic review
-randomized control trial
-cohort study
-case study
-cross sectional study
-case reports
What is prevalence?
the frequency
What is etiology/risk? What studies do you use?
what causes the problem
-cohort study > case control > case series
What is diagnosis? What studies do you use?
whether or not the person has it
-blind comparison to gold standard or cross sectional
What is therapy? What studies do you use?
What is the best treatment
-randomized control study > cohort study
What is prognosis? What studies do you use?
who will get the problem?
cohort study > case control > case series
What is the set up for PubMed?
use PICO for key words during search
-indexed with MeSH terms and newer items not indexed
What is the set up for Cochrane library?
systemic reviews and cochrane reviews
-have their own journals
What is the set up for ADA’s EBD site?
keep up with hot topics; great for once you are out of school
What is the set up for TRIP?
pulls up reviews easily
What is the set up for CEBD?
do not need to use as much and harder to use
What are the limit of databases?
language, age, time period/date, gender/sex. publications types
What is a MeSH term?
medical subject heading
What are Boolean operators?
connection word or symbol that allows a computer user to include or exclude items in a test search
What does OR do in a search?
broaden search
What does AND do in a search?
narrows search
What does NOT do in a search?
excludes
What is a research question?
presents the idea that is to be examined in a study
What is a hypothesis?
testable predication; attempts to answer research question
What is PICO?
a research question that cannot be tested
What is research?
results from individual studies
What is evidence?
accumulation of results
-synthesis of all valid research studies that answer a question
What is validity?
closeness to the truth
What are meta-analyses?
subset of systemic reviews
-method for combining qualitative and quantitative data form several selected studies to develop a single
What are systemic reviews?
provides a comprehensive review of all relevant studies on a particular clinical topic
What are randomized control trials?
participants are randomly assigned to an experimental group or control group
-strongest evidence for cause and effect
-includes at least one varied condition
What are cohort studies?
one or more samples followed prospectively and subsequent status evaluations are conducted
What are case studies?
compare patients who have a disease or outcome of interest with patients who do not and look retrospectively to determine relationships
What are cross-sectional studies?
analysis of data collected from a population of representative subset at one specific point in time
What are case reports?
article that describes and interprets an individual case
-written in form of a detailed story
What is a variable?
unknown or known factors that are relevant to a study
-age, ethnicity, disease history, etc
What is an independent variable?
causastive factor
-factor of conditions that changes naturally or is intentionally manipulated by investigator
What is a dependent variable?
response or outcome
What are confounding variables?
error or confounded influences
-is not considered to have significant impact on dependent variable or outcome
Can you accept a null hypothesis?
NO
what is a P-value?
“alpha value”
-likelihood of differences occurring my chance alone
What does a p-value of equal to or less than 0.05 mean?
reject the null and accept alternative
What test do you use with nominal/categorical data?
Chi square test
What tests do you use for ordinal data?
Mann/Whitnet
What tests do you do for interval data?
T-test or ANOVA
chi square for
nominal data
Mann-whitney for
ordinal data
ANOVA for
interval/ratio with 3+ independent groups
T-test for
interval/ratio with 2 independent groups
inferential statistics do not:
prove cause and effect
-estimate clinical effectiveness
-estimate risk/benefit
Inferential statistics do:
estimate probability of getting results due to chance
-suggest numerical difference
What is needed to calculate effect size?
mean and standard deviation for each group
When relative risk= 1
there was not a difference between groups
When relative risk > 1
risk of bad outcome decrease
When relative risk < 1
risk of bad outcome increased with intervention
What do Odds Ratios compare?
odds of event in one group to odds of events in comparison groups
Sensitivity
negative rules disease out
-can still have false positive
Specificity
positive rules a disease in
-can say you are negative even if you do have the disease
positive predictor value
chance that when a test is positive, you actually have the disease
negative predictor value
chance that when the test is negative, you actually do not have the disease
statistical significance
used in hypothesis testing, may appear to be significantly statistically
-may not affect clincally
clinical significance
practical importance of a treatment effect
experimental event rate
event rate in treated/affected group
control event rate
event rate in the control/unaffected group
absolute risk reduction
compares treatment effectiveness
