Review:

Population: entire group of individuals interested in

Parameters: pop. mean (μ) and pop. standard deviation (σ)

Sample: small group from pop. that is actually tested

Statistics: sample mean (x) and sample standard deviation (s)

Location: mean (average), median (middle), mode (value with highest frequency)

Spread: variance, standard deviation, range

Normal distribution: 689599.7 (important for inference)
Observational Studies vs. Experiments

Observation Study: observes individuals and measures variables of interest but does not attempt to influence the response
(no control, observe sample of pop., not possible to isolate effects of interest, no inference about causation, broader
applications)

Experiment: imposes some treatment on individuals in order to observe their responses (control, impose treatments on
samples, possible to isolate effects of interest, inference about causation, limited to labs/isolated environment)
Introduction to Experimental Design

Planning an experiment: First, form a hypothesis from scientific questions of interest then translate them into specific plans;
Decide on content (response what to measure, treatments conditions to study, units experimental units); Sources of
variability (due to conditions of interest (effects of treatment – wanted), due to the measurement process (measurement
error – don’t want but care can be taken to avoid bias (systematic bias); avoid bias by being careful and keeping conditions
constant), due to experimental material (don’t want but care can be taken to avoid bias (selection bias); randomly assign
treatments to materials to minimize variability); Size of chance errors can be estimated by making several measurements
under the same conditions; To isolate the effects of interest, control what you can and randomize the rest

Three basic principles and four experimental designs (how to assign conditions to material): Random assignment (assigning
condition to material using a chancelike device); Randomizing brings two advantages (protects against possible bias, makes
statistical analysis possible because analysis is based on a set of assumptions); Blocking (subdividing experimental material
into groups/block of similar units then assigning conditions within each block. If variability between experimental units is
known, then block); Factorial crossing (If there are two or more sets of conditions in the same experiment, take all possible
combinations as treatments then randomize or block. Analysis includes interaction effects)

Factor structure of designs: Two universal factors in all designs:
grand average
,
residual error
; structural factors for the
four basic designs:
oneway randomized basic factorial
(randomly assign treatments to units, units are interchangeable;
each treatment is assigned to same number of units – balanced; one structural factor – treatments/conditions);
complete
block
(first sort experimental material into blocks of similar units and then randomly assign treatments to units separately