Biology 2244A/B Chapter Notes -Clinical Trial, Royal Institute Of Technology, Sampling Error
Document Summary
Population the complete collection of all elements (or objects) to be studied: a parameter. Sample: sub collection of members (data points, individuals etc) from a population, a statistic. Interval ordered, differences meaningful, no natural zero. Voluntary response people choose to participate or not: usually people with strong interest/opinion. Observational do not modify subjects, observe and measure, association not causation. Experiment apply treatment, observe effect: ex. Cross-sectional data measured at one point in time. Prospective (cohort/longitudinal) data from future, groups have common factors. Sampling concerns - need a high number of samples due to variability to better represent the entire population (avoid selection bias) Sample size - want to use a sample that is large enough to see the true nature of characteristics of the population. Confounding when effect of variables mix and individual effects can"t be identified. Sampling error diff. between sample result and true population results. Blinding - single (subject) vs. double (subject and experimenter)