Final Flashcards
The extent to which a specific intervention/procedure produces a beneficial result under IDEAL conditions
Efficacy
The extent to which a specific intervention/procedure when deployed in the field in routine circumstances works
Effectiveness
Hierarchy of Evidence
- Systematic reviews and meta-analysis
- Clinical human trial
- Longitudinal cohort studies
- Case-control studies
- Descriptive and cross-sectional studies
- Case reports and case series
- Personal opinion, subjective impressions, anecdotal accounts
Impact Factor
Total number of citations to articles appearing in journal
___________
Total number of articles published
Medical research that follows subjects to assess occurrence of disease (future occurrence). Select based on risk factor. Less bias but more time/money.
Prospective Cohort Study
Groups are defined on the basis of exposure to a suspected _________ for disease
Risk factor
Historical study (looks back on records). Subjects selected based on whether they have disease or not. Efficient in time/cost but susceptible to bias.
Case-Control Study
Snapshot assessment of prevalence of disease/exposure
Cross-sectional Survey
Describe the experience of a single individual/group with similar diagnosis
Case Reports Case Series (multiple CR)
An element, feature, or factor that is liable to vary or change
Variable
A measurement scale based on categorization (gender, political party)
Nominal
A measurement based on relationship between observations (poor-fair-good scale).
Unequal intervals.
Ordinal
Continuous measurement scale of equal units of measurement. (Fahrenheit, Celsius)
Known distance between numbers.
Interval
Continuous measurement scale of equal units with a true zero point at its origin. (Mass, time)
Ratio
Most useful with nominal scale
Mode
Most useful with ordinal scale.
Median
Difference between max and min
Range
A lower and upper confidence bound
Confidence Interval
Average of the square of deviations of measurements about their mean. BEST MEASURE OF SPREAD.
Variance (s2)
Positive square root of the variance. “Natural” variability
Standard Deviation (s or SD)
Measures the percentage of spread
Coefficient of Variation (CV)
SD of sample means. “Error” variability
Standard error (SE)
Used to measure variability to individual subjects around sample mean
SD
Used to asses how accurately a sample mean reflects the population mean
SE
A mathematical statement of no difference. Theory proven when this is rejected.
Null Hypothesis
Proven by showing that the null hypothesis is unlikely to be true
Alternate Hypothesis
Directional Hypothesis
One-tailed
Non-directional hypothesis (states groups are unequaled but no direction difference specified)
Two-tailed
The variable we measure and compare (outcome variable)
Dependent variable
The variable we manipulate (grouping/predictor variable). Used to differentiate groups in a study.
Independent variable
Alpha/p-value error.
Probability of rejecting null hypothesis when it is actually true
Type I error
“Beta error”
Probability of accepting the null hypothesis when it is false
Type II error
The probability of rejecting a null hypothesis when it is false. The ability of the statistical test to detect a specified difference if that difference exists.
Power
Most useful with interval/ratio (continuous) measurement scales
Mean
According to the _______, the sample means of a _____ distributed population will have a ______ distribution
Central Limit Theorem
Non-uniformly
Uniform
A language-based statement of what we are trying to prove.
Research Hypothesis
Types of papers published in journals:
- Research reports
- Reviews of literature (summarization)
- Commentaries
Components of a Research Report:
- Title
- Authors
- Date of submission and acceptance
- Abstract/summary
- Intro
- Materials and Methods
- Results
- Discussion
- Conclusion
The mean is making an _____ from your sample to the entire _____.
Inference
Population
Gives you an idea of how much variation there is in results.
SD
Tells you that if you repeated your experiment many times, the true population mean would be within 1 SD of your sample mean 60% of the time.
SE
As you decrease the amount of error (95 -> 99%) the confidence interval will _____
Increase
How are Alpha and Beta related to each other?
How are they related to the sample size (N)?
They are inversely proportional to each other. (Alpha increases -> beta decreases)
They are inversely proportional to the Sample Size (sample size increases -> alpha and beta decrease)
For nominal/ordinal the best measure of spread is:
Range
Middle 50%
Interquartile Range
Methods to increase power:
- (Inc/Dec) The type 1 error you are willing to tolerate.
- (Inc/Dec) sample size
- (Inc/Dec) deviation from null hypothesis you are willing to tolerate
- (Inc/Dec) variability
- Use a _____ hypothesis if appropriate
- Use the most ______/_____ statistical test
- Increase
- Increase
- Increase
- Decrease
- Directional alternate
- Efficient/Powerful
Statistical Decision-making steps:
- Make a research question
- Define variables
- State null and alternate hypotheses
- Choose statistical test
- Determine type 1 error you will tolerate and sample size required
- Conduct experiment
- Calculate the test statistic
- Determine type 1 error (p-value) using sampling distribution
- Conclude
Describe patterns of disease occurrence in relation to persons, place, and time. This data essential for public health administrators and epidemiologists.
Descriptive Studies
Measures representing characteristics of entire populations are used to describe disease in relation to some factor (age, food, health services)
Correlational Studies
Measures strength of association/relationship
Pearson Coefficient (r)
Indicates change in dependent variable for every one unit change in independent variable
Regression coefficient (B1)
Probability that the results occurred by chance alone and the null hypothesis is really true
P-value
If p-value is less than .05, then reject the null
Indicates value of the dependent variable when the independent variable is 0
Intercept (B0)
Measures the strength of the relationship by “explaining” the % variance of the dependent variable “accounted for” by the independent variable
Coefficient of Determination (R^2)
Correlation Studies strengths and limitations:
Strengths:
- Quick
- Cheap
- Uses available data/info
Limitations:
- Can’t link disease/exposure to individual (data is non-specific)
- Can’t control for effect of potential confounding
- Lack of correlation doesn’t imply an absence of association
- ONLY VALID FOR RANGE STUDIED
Test used to determine if your sample is different from a specified population
Dependent (One-sample) t-test
Divides/partitions the variance of the entire experiment into two or more components
Analysis of Variance (ANOVA)
Represents the average SS for each line
Mean Squares (MS)
Assumptions of ANOVA tables
- Interval/ratio (continuous) measurement scales
- Populations follow normal distribution
- Populations have equal variance
- Independent groups
Theoretical basis for comparisons prior to study. Does not involve testing all possible comparisons. A priori.
Planned Comparisons
Effects suggested by data. A posteriori (post hoc)
Unplanned Comparisons
Single-Variable Tests:
Z-test and T-test
We know the population standard deviation, we estimate the mean from a sample
Z-test
We estimate both the mean and standard deviation
T-test
Dependent Variable- Nominal/Ordinal
Independent Variable- Nominal Ordinal
Test type?
Chi-Square
Dependent Variable- Nominal/Ordinal
Independent Variable- Continuous
Test types?
Logistic Regression
Dependent Variable- Continous
Independent Variable- Continuous
Test types?
Correlation and Regression
Likelihood of developing the disease in the exposed group relative to those not exposed
Relative Risk (RR)
Dependent Variable- Continuous
Independent Variable- Nominal/Ordinal
Test types?
t-Test or ANOVA
The number of people in a population who have a given disease at a given point in time. Frequency of all current cases of disease.
Prevalence
Measure of the number of lesions/period of time. Only measures the numbers of new initial lesions per unit of time.
Incidence
Describes the amount (prevalence) of dental caries in an individual.
DMFT/DMFS
Decayed
Missing
Filled
Teeth/Surfaces
Chi-square tests and Fisher’s Exact tests are similar to _____ tests
Linear Regression and Correlation
Wilcoxon Rank Sum Tests are equivalent to _____ tests
Two-sample t-tests (Independent t-tests)
Kruskal-Wallis Tests are equivalent to _____ tests
ANOVA
Wilcoxon Sign Rank Tests and McNemar tests are equivalent to ______ tests
One-sample t-test (Paired t-test) (Dependent t-test)
Degrees of freedom of Dependent/One-sample t test is equal to
N - 1
N = population size
Degrees of freedom for independent (two-sample) t-test is equal to
Total sample size minus 2
Each individual is measure multiple times
Dependent (one-sample) t-test
Each individual is measured only once and two groups are compared
Independent (two-sample) t-test