Final Exam Review Lecture Flashcards
What is the difference between efficacy and effectiveness?
- efficacy - the extent to which a specific procedure produces a beneficial result under ideal conditions
- effectiveness - measure of the extent to which a specific procedure, when deployed in the field in routine circumstances, does what it is intended to do for a population; “how it works in the real world”
What is a nominal scale of measurement? Give an example.
- process of categorization
- gender, political party, etc.
What is an ordinal scale of measurement? Give an example.
- measurement scale based on the classification of an observation according to its relationship to other observations; numbers not to scale
- poor-fair-good rating, etc.
What is a continuous scale of measurement? Give an example.
- interval or ratio; scale with equal units of measurement
- temperature, mass, time, etc.
What scale of measurement is best described by mode?
nominal
What scale of measurement is best described by mean?
continuous (interval/ratio)
What scale of measurement is best described by median?
ordinal
What is the term for the number of observations that are “free to vary”?
degrees of freedom
What is the difference between standard deviation (SD) and standard error (SE)?
- standard deviation - used to measure the variability of individual subjects around a sample mean
- standard error - used to assess how accurately a sample mean reflects a population mean
Describe the Central Limit Theorem.
In random samples of N observations drawn from a population with a mean (M) and a standard deviation (S), the sample means will be approximately normally distributed with a mean equal to M and a standard deviation equal to (S)/SQRT(N).
What is the purpose of significance?
allows us to make inferences about the population from which our samples are drawn
As sample size increases, error ______.
decreases
What is the difference between a research hypothesis, null hypothesis, and alternative hypothesis?
- research hypothesis - language-based statement of what we are trying to prove
- null hypothesis - mathematical statement of no difference
- alternative hypothesis - mathematical statement that there is a difference; covered everything the null doesn’t
What is Type I error? What is this also called?
- saying there are effects when there are none
- alpha, p-value
What is Type II error? What is this also called?
- saying there is no effects when there are
- beta
What is the relationship between alpha and beta?
inversely proportional
As sample size increases, alpha and beta ______.
decrease
What is power?
- rejecting the null hypothesis when the null hypothesis is false (saying there is a difference when there actually is a difference)
- in other words, the likelihood that you will find a difference if a difference exists
If alpha increases, beta ______, and power ______.
- beta decreases
- power increases
If sample size increases, alpha ______, beta ______, and power ______.
- alpha decreases
- beta decreases
- power increases
Name the hierarchy of evidence quality in order.
- systematic reviews and meta-analysis
- clinical trial in humans (all criteria met)
- clinical trial in humans (most criteria met)
- longitudinal cohort studies
- case-control studies
- descriptive and cross-sectional studies
- case reports and case series
- personal opinions, subjective impressions, and anecdotal accounts
What is the difference between a dependent and independent variable?
- dependent variable - variable we measure and compare
- independent variable - variable we manipulate or the “grouping variable”
What is a one-sample t-test used to determine?
to determine if your sample is different from a specific population
When doing a two-sample t-test, how does the calculation of degrees of freedom change from doing a one-sample t-test?
one-sample t-test:
df = N - 1
two-sample t-test:
df = N1 + N2 -2
What is a two-sample t-test used to determine?
whether the two groups (a control and experimental group, for example) are different
When doing a two-sample t-test, how does the calculation of standard deviation change from doing a one-sample t-test?
one-sample t-test:
SD = SQRT(SS/df)
two-sample t-test:
SD = SQRT((SS1+SS2)/df)
What are the assumptions/requirements of a t-test?
- continuous measurement scale
- samples are drawn from populations with normal distributions
- samples are drawn from populations with equal variances
- the samples are independent
If a t-test fails to demonstrate significance, what 2 reasons could explain this?
- it is the truth (the populations aren’t different)
- the sample size is too small to show a difference
What is the impact factor?
the total number of citations to articles appearing in the journal divided by the total number of articles published
What is a longitudinal cohort study?
group of individuals defined on the basis of exposure to a suspected risk factor of disease; when study starts, subjects are free of disease
What is a case-control study?
subjects selected on the basis of whether they do or do not have a particular disease
What is a descriptive and cross-sectional study?
“snapshot” in time; disease and exposure assessed at the same time in a well-defined group; may not always be possible to distinguish whether the exposure preceded or followed the disease
What is a case report?
describes the experience of a single patient/individual or groups with a similar diagnosis; may lead to formulation of a hypothesis
What is a case study?
collection of individual case reports
What do descriptive studies describe?
describe patterns of disease occurrence in relation to persons, places and time
How is chi square calculated?
(observed - expected)^2/(expected)
What are the assumptions for chi square?
- both groups are composed of nominal or ordinal data
- both groups are independent
- all categories are mutually exclusive
- all expected frequencies are >5
What is the difference between reliability and validity?
- reliability - how consistent or reproducible observations or measurements are
- validity - how closely your observations, measurement instrument or measurements are to the actual or accepted values
What is the reliability coefficient also called?
kappa
What is the equation to calculate the reliability coefficient (kappa)?
K = (PO - PE)/(1-PE)
PO = observed agreement PE = expected agreement
What are the advantages of a case-control study?
- allows study of diseases with long latency periods
- efficient in time and cost
- allows for adequate numbers of diseased and non-diseased individuals to be identified
- evaluation of rare disease
- can evaluate a wide range of potential etiologic exposures and the interrelationships among them
What are the disadvantages of a case-control study?
- disease and exposure have already happened when participant is entered into study
- susceptible to bias
- selection bias
- recall bias (different reporting)
What is a case-control study used for?
- test a specific hypothesis
- explore a range of exposures among affected or non-affected individuals
What are the three types of cohort study?
- prospective
- retrospective
- both prospective and retrospective
What is the definition of relative risk (risk ratio)? What is its equation?
- likelihood of developing the disease in the exposed group relative to those not exposed
- RR = (incidence in exposed group)/(incidence in non-exposed group)
What type of study is the risk ratio associated with? What type of study is the odds ratio associated with?
- risk ratio - cohort study
- odds ratio - case-control study
What are the advantages of a cohort study?
- temporal sequence between exposure and disease established
- assessing the effects of rare exposures
- adequate numbers of exposed and nonexposed individuals identified
- allows for examination of multiple effects of a single exposure
- minimize potential for selection bias
What are the disadvantages of a cohort study?
- follow-up for many years
- time-consuming
- expensive
- bias associated with loss to follow-up
What is a retrospective cohort study?
exposure and outcome/disease have already happened at the start of the study
What is the difference between cohort and case-control studies?
- cohort - select individuals based on exposure and studying disease
- case-control - select individuals based on disease and studying exposure
- *these are my own definitions
What does an analysis of variance (ANOVA) tell us?
whether the variance is due to treatment effects or due to error
When completing an ANOVA table, how is the df for treatment found?
(# of groups) - 1
When completing an ANOVA table, how is the SS (sums of squares) for treatment found?
SStx = SUM(n(groupmean - grandmean)^2)
n = number of samples in each group grandmean = mean of all groups combined
When completing an ANOVA table, how is the df for error found?
(total number of samples) - (number of groups)
When completing an ANOVA table, how is the MS for treatment or error found?
MS = SS/df
When completing an ANOVA table, how is the F ratio found?
F = MStx/MSerr
What does the F ratio represent?
the probability that the variance between the groups is due to a difference between the groups is times greater than the variance being due to error
What are the assumptions made during a ANOVA test?
- interval/ratio (continuous) measurement scales
- populations follow a normal distribution
- populations have equal variance
- independent groups
What are 2 tests for unplanned comparisons?
- Tukey test
- Dunnett’s
If the independent variable is nominal/ordinal and the dependent variable is nominal/ordinal, what type of test is performed?
chi square
If the independent variable is nominal/ordinal and the dependent variable is continuous, what type of test is performed?
t-test or ANOVA
If the independent variable is continuous and the dependent variable is continuous, what type of test is performed?
correlation and regression
What is the nonparametric equivalent to a linear regression and correlation?
chi square
What is the nonparametric equivalent to a two-sample t-test?
Wilcoxon Rank Sums test
What is the nonparametric equivalent to an ANOVA?
Kruskal-Wallis test
What are the nonparametric equivalent to one-sample t-test?
- Wilcoxon Sign Rank test
- McNemar test
How does a Wilcoxon Rank Sums test work? Explain it.
- arrange all of the data in ascending order
- give a rank so that the smallest number is ranked 1 and so on
- now categorize the data by its nominal group and sum up the ranks of the data in those groups so that each group has a rank sum
- if the null hypothesis is true, each group should have the same rank sum
What is the Kruskal-Wallis test used for?
- extension of the Wilcoxon Rank Sum test
- with two groups, it is like the t-test and ANOVA
- can handle more than two groups
How does the Wilcoxon Sign Rank test work? Explain it.
- handles repeated measures
- take the difference between the repeated measures and place it in order of smallest to biggest difference
- assign each difference a rank
- assign each difference a sign based on whether the difference is +/-
- if the null hypothesis is true, the average for the + and the - ranks should be the same
What type of information is more important in a McNemar test: when the two outcomes have the same result or when they differ?
when they differ because that refutes the null hypothesis
What is the purpose of a diagnostic test?
diagnostic tests indicate the presence or absence of a disease and are usually compared to a “gold standard”
What is the difference between sensitivity and specificity?
- sensitivity - proportion with true disease that is diagnosed as having disease (TP/(TP+FN))
- specificity - ability of test to correctly identify those without disease (TN/(TN+FP))
Which (specificity or sensitivity) rules out the presence of disease? Which rules in the presence of disease?
- specificity rules in the presence of disease
- sensitivity rules out the presence of disease
What is the trade-off between sensitivity and specificity?
need both to be high, but usually one is increased at the expense of the other