Lecture 05a - Epid 301 Precision and Validity Note: This is...

Info iconThis preview shows pages 1–7. Sign up to view the full content.

View Full Document Right Arrow Icon
Epid 301 Precision and Validity Note: This is the slide deck I didn’t show, while several concepts were covered in class and all concepts are in book. View an epidemiologic study as an exercise in measurement. The specific aims involve the: - estimation of event occurrence(i.e. incidence rates) - effect of exposure factors on health Goal is to estimate the parameter of interest with little error - sources of error may be classified as either random or systematic - principles of study design emerge from consideration of approaches to reducing both types of error Random Error - Precision (p-values, confidence intervals, study size and power) Systematic Error - Internal Validity (bias, confounding) - External Validity (generalizability) Feb 11 2010 ARONSON
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Precision Session Objectives : Gain an epidemiologic understanding of the role of chance. Introduce hypothesis testing Statistical significance Power Gain an understanding what parameters of study design influence statistical power Be able to calculate power in the case-control setting.
Background image of page 2
Precision - Lack of random error Inferences are based on a sample Conceptually the subjects in an epidemiologic study are always considered a sample –> Random error from sample to sample Tests of significance Degree to which chance variability may account for observations p-value Probability that an effect as extreme as that observed could be observed by chance (assuming that there is truly no association) Determined by magnitude of effect and sample size
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Testing and Estimation Usually, in epidemiologic studies, we wish to measure the difference in disease occurrence between groups exposed and not exposed to a particular factor Estimate the size of the association (i.e. Relative Risk) and whether an association this large is likely to have been observed by chance (statistical significance) Involves an explicit statement of the hypothesis to be tested: Null Hypothesis (H0): Alternative hypothesis (HA): p0 = p1 p0 p1 OR = 1.0 OR 1.0 RR = 1.0 RR 1.0 means = 0 means 0 (note, hypothesis testing begins with the assumption the null hypothesis is true)
Background image of page 4
Consider the following example: A case-control study examining the relationship between a history of sunburn and risk for melanoma. The following table summarizes the data. History of sunburn Cases Controls Yes 45 32 No 55 68 What is the measure of effect ? What is the hypothesis test ?
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
45*68 = 1.74 32*55 History of sunburn Cases Controls Yes 45 32 No 55 68 Measure of effect ? Test of statistical significance: For the 2x2 table generated above we could calculate a chi square test. X 2 = 3.56, p = 0.06 P-value represents the probability of observing a result at least as extreme as that observed by chance alone (assuming that there is truly no association). It quantifies the degree to which sampling variability may account for the
Background image of page 6
Image of page 7
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 36

Lecture 05a - Epid 301 Precision and Validity Note: This is...

This preview shows document pages 1 - 7. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online