F20-ExpDesigns

F20-ExpDesigns - PubH 7405: REGRESSION ANALYSIS ONE-FACTOR...

Info iconThis preview shows pages 1–9. Sign up to view the full content.

View Full Document Right Arrow Icon
PubH 7405: REGRESSION ANALYSIS ONE-FACTOR EXPERIMENT DESIGNS
Background image of page 1

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

View Full DocumentRight Arrow Icon
Designed experiments are conducted to “demonstrate” a cause-and-effect relation between one or more explanatory factors (or predictors) and a response variable. The demonstration of a cause-and-effect relationship is accomplished, to put it in a simple way, by altering the level or levels of the explanatory factors (i.e. “designed”) and observing the effect of the changes (i.e. designed values of predictors X’s) on the response variable Y. Designed experiments are often used as comparative ” in natures.
Background image of page 2
A Simple Example : An experiment on the effect of Vitamin C on the prevention of colds could be simply conducted as follows. A number of n children (the sample size) are randomized; half were each give a 1,000-mg tablet of Vitamin C daily during the test period and form the “experimental group”. The remaining half , who made up the “control group” received “placebo” – an identical tablet containing no Vitamin C – also on a daily basis. At the end, the “Number of colds per child” could be chosen as the outcome/response variable, and the means of the two groups are compared. Pay attention to the Explanatory variable (Predictor), Factor levels or treatment arms, Experimental units, and Outcome/Response variable.
Background image of page 3

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

View Full DocumentRight Arrow Icon
Assignment of the treatments (factor levels: Vitamin C or Placebo) to the experimental units (children) was performed using a process called “randomization”. The purpose of randomization was to “balance” the characteristics of the children in each of the treatment groups, so that the difference in the response variable, the number of cold episodes per child, can be rightly attributed to the effect of the predictor – the difference between Vitamin C and Placebo.
Background image of page 4
The simplest form of designed experiments is the completely randomized design ” where treatments are randomly assigned to the experimental units – regardless of their characteristics. This design is most useful when the experimental units are relatively homogeneous with respect to known confounders. Otherwise, heterogeneous experimental units are divided into homogeneous “block”; and randomizations of treatments are carried out within each block. The result would be a “ randomized complete block design ”; the analyses are different, say, One-way ANOVA versus Two-way ANOVA .
Background image of page 5

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

View Full DocumentRight Arrow Icon
INFERENCES & VALIDITIES Two major levels of inferences are involved in interpreting a study The first level concerns Internal validity ; the degree to which the investigator draws the correct conclusions about what actually happened in the study. The second level concerns External Validity (also referred to as generalizability or inference ); the degree to which these conclusions could be appropriately applied to people and events outside the study.
Background image of page 6
Truth in The Universe Truth in The Study Findings in The Study Research Question Study Plan Study Data External Validity Internal Validity
Background image of page 7

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

View Full DocumentRight Arrow Icon
With the goal of maximizing the validity of the
Background image of page 8
Image of page 9
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 11/21/2011 for the course PUBH 7405 taught by Professor Staff during the Fall '08 term at Minnesota.

Page1 / 58

F20-ExpDesigns - PubH 7405: REGRESSION ANALYSIS ONE-FACTOR...

This preview shows document pages 1 - 9. Sign up to view the full document.

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