Biostats- Melissa Wells Flashcards
What is the Belmont Report?
Belmont report = basic ethical principles underlying proposed regulations for research
What are the three principles underlying the Belmont Report?
- Respect for persons (informed consent)
- Beneficence (assessment of potential risks and benefits)
- Justice (selection of people to be in the research)
What can be determined from an observational study?
Associations or correlations
NO CAUSALITY
What is an observational study?
a type of study in which individuals are observed or certain outcomes are measured. No attempt is made to affect the outcome (no treatment is given)
What is a cross sectional study?
- a type of observational study
- one point in time where independent and dependent variables are measured at the same time
What are the three types of longitudinal studies? and how are they different?
- Trend- measure changes in population over time (ex: a survey of a college freshman each year)
- Cohort- measures change and follows a particular population over time (not necessarily the SAME people though)- prospective (followed for longitudinal time period)
- Panel- measure changes in the SAME people over time
Qualitative study characteristics and some examples
- High validity
- analysis of less structured data (open ended response, etc)
- Focus groups, qualitative personal interviews, visual methodology
Quantitative characteristics
- High reliability
- Often coupled with generalizable sampling procedures
Mixed methods include…
both qualitative and quantitative aspects of study
Type of nomothetic causation
- Correlation/Association
- Temporal Order (A before B)
- Non-spuriousness (correlation, but not associated)
What are the three main parts of the “classical experiment”?
- Independent and dependent variable
- Pre-testing & Post-testing
- Control and experimental groups
What is investigated in phase I of clinical trials?
- Safety
- Side-effects
- Metabolism
Note: usually done on healthy volunteer, check metabolism in actual person
What is investigated in phase II of clinical trials?
- Ideal dosing
Note: Need patients with condition
What is investigated in phase III of clinical trials?
- Experimental vs. control
“Hallmark of Clinical Trials”
What is investigated in phase IV of clinical trials?
Continued evaluation of FDA-approved therapy (Post-market study)
What is the Hawthorne effect?
When researcher behavior affects behavior
What is a double blind study?
research and respondents don’t know which treatment is the control or experimental
What is external invalidity?
Applicability to real world, the sample population isn’t really representative of the real world.
What is internal invalidity?
anything other than the stimulus that influences results (history, change in maturation, testing, instrumentation, etc)
What is a case-control study?
- type of observational study
- a comparison of people who have a condition and/or receive a treatment (cases) with another group of people who are not affected by the condition (control)
- needs to be RETROSCPECTIVE
Four methods/types of non-probability sampling?
- Purposive
- Convenience
- Snowball
- Quota
What is a good way to avoid sampling bias??
Use a probability sample (a random sample)
What is the goal of probability sampling?
to a create a sample as REPRESENTATIVE of the population as possible
Element vs. population
element = individual members of population
Population
entire set of elements
Sampling frame
list of all elements in a population
Parameter vs. statistic?
Parameter = summary of given variable in population Statistic= summary of given variable in a sample
What is a sampling distribution?
all possible random samples that could be selected
What are the four types of random sampling?
- Simple random sample
- Systematic
- Stratified
- Multi-stage Cluster (or cluster sampling)
Explain simple random sample?
need a list
assign a numer
select by a random number generator
Explain systematic random sample
determine sample size divide population by sample number (called sampling interval L) list and number elements RANDOMLY SELECT START POINT the select every k-th element with-in groups
Explain stratified random sample
random sample from sub populations
stratify your list
sample with-in those strata
(this increases sampling error potential)
What is sampling error?
variation in values of your sample mean compared to the population mean
What are two ways to reduce sampling error?
- increasing sample size
2. increasing homogeneity
What are some characteristics of the normal curve?
- theoretical distribution of scores
- perfectly symmetrical
- bell shaped
- Unimodal
- Tails extend infinitely in both directions
- Mean/Median/Mode are equal in center of curve
What area of the curve falls within 1 standard deviation of the mean?
68%
What area of the curve falls within 2 standard deviations of the mean?
95%
What area of the curve falls within 3 standard deviations of the mean?
98%
What is the critical z-score of a 95% confidence level? (when alpha = .05)
1.96
What is the critical z-score of a 90% confidence level? (when alpha =.10)
1.65
What is the critical z-score of a 99% confidence level? (when alpha = .01)
2.58
What is confidence interval and how do you find it?
- range within “true” parameters should lie
- CI= mean +/- (z-score x SE)
How do you find the standard error when given the standard deviation?
SE = SD/ (sq rt of sample size)
What is the z-distribution?
a special case of normal distribution where idealized mean is 0 and s.d. is 1
What is the confidence/significance level?
probability that our sample statistic will with with-in a given confidence interval
- we set this ahead of time= alpha
- when alpha is = .05 we are 95% sure that the statistic will fall within the given confidence interval
What can influence confidence intervals?
- Confidence level (alpha can be lowered or raised)
- Sample size (more confidence with larger samples, so smaller interval)
- Variation: more variation= more error (averages, high standard deviations)
Define hypothesis
a prediction about the relationship between 2 variables that asserts that changes in the measure of an independent variable will correspond to the changes in the measure of the dependent variable
Research hypothesis (H1)
predicts differences and relationships
Null Hypothesis (H0)
predicts no relationship or no difference
-WE TEST THE NULL HYPOTHESIS
What is a test statistic?
a measure of how different finding is from what is expected from the null hypothesis
What is the z-distribution?
- to use normal curve to answer questions, raw scores of distribution must be converted into z-scores
- critical regions are areas under the curve that include unlikely sample outcomes if the null hypothesis is true
- z-critical establishes critical regions (size of this region is reported by alpha)
Do you reject H0 if the test statistics falls into the critical region?
YES!
What happens if the test statistic does NOT fall into a critical region?
FAIL TO reject null hypothesis
What does a p-value tell you?
The probability of obtaining this finding if the null is correct
p > alpha
we do NOT reject the null hypothesis
p < or = alpha
REJECT the null hypothesis
-P-value is STATISTICALLY SIGNIFICANT
Pearson’s correlation coefficient
-1 to 1
slopes
(closer to -1 or 1 the stronger the correlation)
Type 1 error
- rejecting the null when the null is true
- False positive
- Saying the are different, when they really are not
Type II error
- failing to reject the null when the null is false
- False negative
- Saying the are the same, when really they are not
What is confounding variable?
A third variable that truly accounts for the relationship of two other variables, A and B
(ice cream sales correlates with crime rates - but one doesn’t cause that other, there is a confounding variable - summer
-A and B relationship considered spurious (no relation)