design of experiments

Summary

Changes of univariate summary statistics, and bivariate r and m

Var - variance, average suqared deviation from the mean
SD - standard deviation, spread about the mean
COV - coefficient of variation, degree of spread relative to mean

Random Assignment -> proportional distribution of different characteristics between treatment and control, to reduce confounding
Probability Sampling -> probability of being selected is known and non-zero

Concept

Types of variables

• Categorical Nominal
• Categorical Ordinal
• Numerical Discrete
• Numerical Continuous

Probability sampling methods

• Simple Random
• Systematic (interval)
• Stratified Random (proportion from each strata)
• Cluster

Consequencs of non-probability sampling

• Selection Bias
• Non-response Bias

Criteria for Generalisability

• Sample Frame >= Target Pop
• Probability Sampling
• Sufficient Response Rate
• Accurate Data

Experimental Studies

• Treatment vs Control groups
• Blinding
  • Placebo
  • Double-blinding, both accessors and participants

Observational Studies

• Self-selecting
• Only provides evidence of association