Quotes of the day

1
Quotes of the day

• Prediction is very difficult, especially about
  the future.
• Niels BohrDanish physicist (1885 - 1962)
• When a distinguished but elderly scientist states
  that something is possible, he is almost
  certainly right. When he states that something is
  impossible, he is very probably wrong.
• Arthur C. Clarke, Clarke's first lawEnglish
  physicist science fiction author (1917 - )

2
What is hypothesis testing?

• Hypothesis-driven research
• A good hypothesis is one that can be tested
• A good experiment is one that unequivocally tests
  an hypothesis
• Most experiments are based on quantitative data.
  These require statistics for analysis

3
Biologists and Statistics

• The drunk and the lamp post
• Suspicion How to lie with statistics
• Misuse gives stats a bad name
• Misuse of stats inappropriate test gives
  invalid answers
• Strategic planning in retrospect
• Statistical analysis should be part of
  experimental design.

4
Why statistics some definitions

• Populations vs samples
• We rarely have the opportunity to assess a
  parameter in the entire population
• We sample the population and try to infer from
  the sample what would happen in the population.
• Example Interleukin 12 response in virus
  infected mice. We cant measure the response in
  all mice, so we sample the population.
• Eventually we try to extrapolate to relevant
  species.

The fundamental question How do we know if an
unknown sample is part of the population or if it
is part of a different population?
5
Population parameters
N 5
8
6
The sums of the differences between individual
observations and the mean is always 0.
How can we come up with a number that estimates
how variable the data are?
7
The miracle of the square.
Variance
Standard Deviation
8
Now lets sample that population
9
Random sampling involves error in estimating the
population
10
(No Transcript)
11
Population terms
12
What is an experiment?

• Types of studies
• Descriptive No experiment observation
• Correlation May need experiment
• Causality Always requires experiment
• To design an experiment, first need research
  hypothesis.
• This is different from statistical hypothesis

13
Variables

• Types
• Independent
• Dependent
• Extraneous (nuisance) variables
• Classifications of variables
• Measurement variables
• Continuous
• Discontinuous
• Ranked variables (e.g. development stages)
• Attributes or nominal variables
• Typically collected as a frequency of occurrence.

14
Sampling a population

• Problems with generalization
• Sampling procedure Random sampling and
  allocation
• Statistical analysis assumes random sampling
• Must distribute nuisance variables randomly among
  groups- avoid systematic error.

15
Basic structure of experimental designs

• Between subjects
• Within Subjects
• Factorial

16
Statistical hypothesis testing

• The null hypothesis
• States that groups are sampled from the same
  population
• Statistical analysis either accepts or rejects
  the null hypothesis.
• What do we mean by p lt 0.05?

17
Types of error

• Type I reject a true null hypothesis
• Type II accept a false null hypothesis