Epidemiology Methods and Biostatistics Flashcards
?: extent at which an instrument measures what it is supposed to measure
Validity
?: general accuracy of the measure
Validity
4 types of validity
- Face validity
- Content validity
- Construct validity
- Criterion validity
?: is the question/ first look/ overall structure valid at first glance
Face validity
?: does it measure what it is intended to measure
Content validity
?: ability to measure abstract concept adequately, it it designed to measure well
Construct validity
?: extend to which measurement able to produce accurate findings when compared to gold standard
Criterion Validity
?: is degree to which a measurement produces consistent outcomes
Reliability
3 types of reliability:
1- Test-restest reliability
2- internal consistency
3- inter-rater reliability
?: refers to repeatability. Such as food frequency questionnaire
Test-restest reliability
?: similar questions give similar answers to same subject
Internal consistency
?: looks at consistency and agreement between two experts; same questions by two different administrators
inter-rater reliability
____ is related to general accuracy
Validity
____ is related to general precision
Reliability
____ is related to validity and accuracy
Bias
____ is related to reliability and precision
Random Error
Bias is:
systematic error
Precision is:
random error
Types of Bias
- Observer Effect
- Social Desirability Bias
- Selection Bias
- Sampling error
- Reporting bias
- Publication bias
- Measurement error
- Misclassification error
- Nonresponse bias
- Healthy volunteer bias
?: researcher or interviewer body language or intonation affects participant’s response
Observer Effect
?: when participant responds in a way they think they should
Social desirability bias
?: ppl decide to include in study dont represent population aimed to study. MAIN THREATS to VALIDITY AND GENERALIZABILITY
Selection bias
?: when data points used in analysis don’t represent entire population of data
Sampling error
?: direction or statistical significance of results influence whether or how research is reported
Reporting bias
?: when papers finding a null hypothesis are not submitted to review, or or not selected for publication
Publication bias
?: random error, miscalibrated instrument, imprecision
Measurement error
?: wrong category placement
Misclassification error
?: not enough answers or data collection when send out surveys (1000 send vs 100 responds)
Non-response bias
?: healthy ppl participate in healthy surveys
Healthy volunteer bias
?: how valid and close to the truth
Accuracy
?: having frequent similar results not necessarily close to truth
Precision
?: the ability of a test to correctly identify those with a disease
Sensitivity
?: ability of a test to correctly identify those without a disease as disease free
Specificity
Sensitivity formula
TP/ TP + FN
Specificity Formula
TN/ FP + TN
?: probability that a pt has the disease given a positive screening test value
Positive Predictive Value PPV
?: probability pt does not have disease given negative screening test value
Negative predictive value NPV
PPV Formula
TP/ TP + FP
NPV Formula
TN/ FN + TN
?: number of existing disease cases during a period of time
Prevalence
?: number of new disease cases that develop over a period of time
Incidence
PPV increase with increasing ____
Prevalence
?: state of being unhealthy, not in good health, or with disease
Morbidity
?: refers to death
Mortality
?: measure of association between a dependent and independent variable
Odds ratio
?: average across a population
Mean
?: 50th percentile
Median
?: most frequent value
Mode
With a negative skew, ____is lower than ____
Mean is lower than median
With positive skew, ____is higher than ____
Mean is higher than median
?: way to describe spread of a distribution
Standard deviation