Advanced Statistics

Advanced Statistics
Janette Walde
[email protected]
Department of Statistics
University of Innsbruck
Advanced Statistics

Contents
Introduction
Basics/Descriptive Statistics
Scales of measurement
Graphical exploration of data
Descriptive characteristics for a variable
Estimation
Characteristics of an estimator
Confidence interval
Statistical hypothesis testing
Statistical testing principle
Testing errors
Power analysis
Why multivariate analysis?
Advanced Statistics
Introduction

“We are pattern-seeking story-telling animals.”

(Edward Leamer)

”Statistics does not hand truth to the user on a

silver platter. However, statistics confines
arbitrariness and provides comprehensible
conclusions.”

“Es gibt keine Tatsachen, es gibt nur

Interpretationen.” (Friedrich Nietzsche)
Advanced Statistics
Introduction

Preliminary comments
1. You will learn to apply statistical tools correctly,
interpret the findings appropriately and get an
idea about the possibilities of analyzing
research questions employing statistics.
2. It is not possible and not worthwhile to learn
all statistical methods in such a course.
However, this course is successful if it enables
you to improve your knowledge in statistical
methods on your own. Therefore this course
gives you profound knowledge about some
statistical analyzing tools and shows you the
correct application of them.
Advanced Statistics
Introduction

Preliminary comments
3. Although knowing the most sophisticated
analyzing instruments one may be confronted
with limits in getting results or finding
appropriate interpretations or applying tools in
the given framework. This has to be accepted
(“If we torture the data long enough, they will
confess.”).
4. Be aware: Never confuse statistical significance
with biological significance.
Advanced Statistics
Basics/Descriptive Statistics
Scales of measurement

Scales of measurement
1. Nominal Scale. Nominal data are attributes like
sex or species, and represent measurement at
its weakest level. We can determine if one
object is different from another, and the only
formal property of nominal scale data is
equivalence.
2. Ranking Scale. Some biological variables
cannot be measured on a numerical scale, but
individuals can be ranked in relation to one
another. Two formal properties occur in
ranking data: equivalence and greater than.
Advanced Statistics
Basics/Descriptive Statistics
Scales of measurement

Scales of measurement

3. Interval and Ratio Scales. Interval and ratio

scales have all the characteristics of the ranking
scale, but we know the distances between the
classes. If we have a true zero point, we have a
ratio scale of measurement.
Advanced Statistics
Basics/Descriptive Statistics
Graphical exploration of data

Histogram
Normal distribution Skewed distribution
300 300

250 250

200 200
frequency (density)

frequency (density)
150 150

100 100

50 50

0 0
−4 −3 −2 −1 0 1 2 3 4 0 2 4 6 8 10 12 14 16 18 20
X Y
Advanced Statistics
Basics/Descriptive Statistics
Graphical exploration of data

Box Plot
Normal distribution Skewed distribution

3 18

16
2
14
1

frequency (density)
frequency (density)

12

0 10

8
−1
6
−2
4

−3 2

0
−4
X Y
Advanced Statistics
Basics/Descriptive Statistics
Graphical exploration of data

Q-Q Plot
I Many statistical methods make some
assumptions about the distribution of the data
(e.g. normality).
I The quantile-quantile plot provides a way to
visually investigate such an assumption.
I The QQ-plot shows the theoretical quantiles
versus the empirical quantiles. If the
distribution assumed (theoretical one) is indeed
the correct one, we should observe a straight
line.
Advanced Statistics
Basics/Descriptive Statistics
Graphical exploration of data

Q-Q Plot
Normal Q−Q Plot Normal Q−Q Plot

50
2

40
1
Sample Quantiles

Sample Quantiles

30
0

20
−1

10
0
−2

−2 −1 0 1 2 −2 −1 0 1 2

Theoretical Quantiles Theoretical Quantiles

0.4
0.3
density

0.2
0.1
0.0

−4 −2 0 2 4
Advanced Statistics
Basics/Descriptive Statistics
Descriptive characteristics for a variable

Summary Statistic

I Mean, median
I Percentiles, inter quartile range
I Minimum, maximum, range
I Standard deviation, variance
I Coefficient of variation
I Median absolute deviation, mean absolute
deviation
Advanced Statistics
Estimation

Fundamental concepts
Populations must be defined at the start of any
study and this definition should include the spatial
and temporal limits to the inference. The formal
statistical inference is restricted to these limits.

Possibility of drawing samples randomly.

Population parameters are considered to be fixed

but unknown values (in contrast to the Bayesian
approach).
Advanced Statistics
Estimation
Characteristics of an estimator

Characteristics of an estimator
A good estimator of a population parameter should
have the following characteristics:
I The estimator should be unbiased, meaning

that the expected value of the sample statistic

(the mean of its probability distribution) should
equal the parameter.
I It should be consistent so as the sample size

increases then the estimator will get closer to

the population parameter.
I It should be efficient, meaning it has the lowest

variance among all competing estimators.

Advanced Statistics
Estimation
Characteristics of an estimator

Unbiasedness of sample mean as estimator

for the population mean
n = 50
0.3

0.2

0.1
mean of each sample

−0.1

−0.2

−0.3

−0.4
1 2 3 4 5 6 7 8 9 10
number of sample
Advanced Statistics
Estimation
Characteristics of an estimator

Consistency of the sample mean as

estimator for the population mean
n = 10

−5

1 2 3 4 5 6 7 8 9 10

n = 100

−5

1 2 3 4 5 6 7 8 9 10

n = 10,000

−5

1 2 3 4 5 6 7 8 9 10
Advanced Statistics
Estimation
Characteristics of an estimator

Efficiency of the sample mean and of the

median as an estimator for the population
central tendency
1,000 samples with n = 100, variabe is normally distributed with population mean zero and standard deviation ten
4

2
distribution of the means

−1

−2

−3

−4

−5
mean estimator median
Advanced Statistics
Estimation
Confidence interval

Confidence interval for the population

mean
Consider a population of N observations of the
variable X . We take a random sample of n
observations {x1, x2, ..., xn} from the population.
I Median versus sample mean (x̄).

I Having an estimate of a parameter is only the

first step in estimation. We also need to know

how precise our estimate is: Standard error.
Standard error of the mean: sex̄ = √σ̂n
I Confidence interval for the population mean:

CI(1−α) : [x̄ − tdf =n−1,1−αsex̄ ; x̄ + tdf =n−1,1−αsex̄ ]

Advanced Statistics
Estimation
Confidence interval

95% confidence interval for the population

mean
n = 10
10

−5

−10
1 2 3 4 5 6 7 8 9 10

n = 100
10

−5

−10
1 2 3 4 5 6 7 8 9 10

n = 10,000
0.4

0.2

−0.2

−0.4
1 2 3 4 5 6 7 8 9 10
Advanced Statistics
Statistical hypothesis testing
Statistical testing principle

Statistical tests and scientific hypotheses

A statistical test is a confrontation of the real world

(observations) to a theory (model) with the aim of
falsifying the model.

Model: H0 : µ = 0 and Ha : µ 6= 0
Real world: x̄, s
Advanced Statistics
Statistical hypothesis testing
Statistical testing principle

Statistical tests and scientific hypotheses

As such the statistical test (as a scientific method)
fits directly into the philosophy of science described
by the English philosopher Karl Popper (1902–1994)
(see e.g. The Logic of Scientific Discovery, 1972).
Basically the philosophy says that 1) theories can
not be empirically verified but only falsified and 2)
scientific progress happens by having a theory until
it is falsified. That is, if we observe a phenomenon
(data) which under the model (theory) is very
unlikely, then we reject the model (theory).
Advanced Statistics
Statistical hypothesis testing
Statistical testing principle

Statistical tests and scientific hypotheses

”No amount of experimentation can ever prove me right; a
single experiment can prove me wrong.” (Albert Einstein)

In other words, experiments can mainly be used for

falsifying a scientific hypothesis – never for proving
it! When we have a scientific theory, we conduct an
experiment in order to falsify it. Therefore, the
strong conclusion arising from an experiment is
when a hypothesis is rejected. Accepting (more
precisely – not rejecting) a hypothesis is not a very
strong conclusion (maybe acceptance is simply due
to that the experiment is too small).
Advanced Statistics
Statistical hypothesis testing
Statistical testing principle

Example
Suppose we have a coin, and that our hypothesis is
that the coin is fair, i.e. that P(head) = P(tail) =
1/2. Suppose we toss a coin n = 25 times and
observe 21 heads. The probability of actually
observing these data under the model is P(21 heads,
4 tails) = 0.0004. It is a very unlikely (but possible)
event to see such data if the model is true. In this
falsification process we employ the interpretation
principle of statistics:

Unlikely events do not occur...

Advanced Statistics
Statistical hypothesis testing
Statistical testing principle

Statistical tests and scientific hypotheses

If we do not employ this principle we can never say

anything at all on the basis of statistics
(observations): An opponent can always claim that
the present observations just are “an unfortunate
outcome” which - no matter how unlikely they are -
are possible.
Advanced Statistics
Statistical hypothesis testing
Statistical testing principle

Statistical tests and scientific hypotheses

In practice the statistical interpretation principle
needs more structure:
I In a large sample space, all possible outcomes

will have a very small probability, so it will be

unlikely to have the data one has.
I In addition there is also the question about

how small a probability is needed in order to

classify data as being unlikely.
I Concepts of p-value and significance level α.
Advanced Statistics
Statistical hypothesis testing
Testing errors

Two Types of Errors

Recall that the following four outcomes are possible
when conducting a test:

Reality Our Decision

H0 Ha
√
H0 Type I Error
(Prob = 1 − α) Prob = α
√
Ha Type II Error
Prob = β (Prob = 1 − β)
The significance level α of any fixed level test is the
probability of a Type I error.
Advanced Statistics
Statistical hypothesis testing
Testing errors

Acceptable levels of errors

I Type I error (α)
I Typically α = 0.05 (This convention is due to R.A.
Fisher)
I For more stringent tests α = 0.01 or α = 0.001
I Exploratory or preliminary experiments α = 0.10

I Type II error (β)

I Typically 0.20
I Often unspecified and much less than 0.20

I Statistical power=(1 − β)
Advanced Statistics
Statistical hypothesis testing
Power analysis

The power of a statistical test

The power of a significance test measures its ability

to detect an alternative hypothesis.
The power against a specific alternative is
calculated as the probability that the test will reject
H0 when that specific alternative is true.
Advanced Statistics
Statistical hypothesis testing
Power analysis

Example: Computing statistical power

Does exercise make strong bones?

Can a 6-month exercise program increase the total body bone

mineral content (TBBMC) of young women? A team of
researchers is planning a study to examine this question.
Based on the results of a previous study, they are willing to
assume that σ = 2 for the percent change in TBBMC over the
6-month period. A change in TBBMC of 1% would be
considered important, and the researcher would like to have a
reasonable chance of detecting a change this large or larger.
Are 25 subjects a large enough sample for this project?
Advanced Statistics
Statistical hypothesis testing
Power analysis

Example (cont.)
1. State the hypotheses: let µ denote the mean
percent change:
H0 : µ = 0
Ha : µ > 0
2. Calculate the rejection region: The z test
rejects H0 at the α = 0.05 level whenever:
x̄ − µ0 x̄
z= √ = √ ≥ 1.645
σ/ n 2/ 25
That is we reject H0 when x̄ ≥ 0.658.
Advanced Statistics
Statistical hypothesis testing
Power analysis

Example (cont.)
3. Compute the power at a specific alternative:
The power of the test at alternative µ = 1 is
P(x̄ ≥ 0.658|µa = 1) = 0.8
Plot graph.
4. Statistical power is the probability of rejecting
H0 given population effect size (ES), α and
sample size (n). This calculation also requires
knowledge of the sampling distribution of the
test statistic under the alternative hypothesis:
Power curve.
Advanced Statistics
Statistical hypothesis testing
Power analysis

Example (cont.)
Power function in dependence of the effect size
1

0.9

0.8

0.7

0.6
power=1−β

0.5

0.4

0.3

0.2

0.1

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
µ −µ
0 a
Advanced Statistics
Statistical hypothesis testing
Power analysis

Ways to increase power

I Increase α. A 5% test of significance will have
a greater chance of rejecting the alternative
than a 1% test because the strength of
evidence required for rejection is less.
I Consider a particular alternative that is farther
away from µ0 . Values of µ that are in Ha but
lie close to the hypothesized value µ0 are harder
to detect than values of µ that are far from µ0 .
I Increase the sample size. More data will
provide more information about x̄ so we have a
better chance of distinguishing values of µ.
Advanced Statistics
Statistical hypothesis testing
Power analysis

Ways to increase power

I Decrease σ. This has the same effect as

increasing the sample size: it provides more
information about µ. Improving the
measurement process and restricting attention
to a subpopulation are two common ways to
decrease σ.
Advanced Statistics
Statistical hypothesis testing
Power analysis

How many samples are needed to achieve

a power of 0.8 in a t-test?
Effect size index for the t-test for a difference
between two independent means.
µ1 −µ2
d= σ

where d is the effect size index, µ1 and µ2 are

means, σ is the common standard deviation of the
means.

Effect size indices are available for many statistical

tests.
Advanced Statistics
Statistical hypothesis testing
Power analysis

How many samples are needed to achieve

a power of 0.8 in a t-test?
Effect Size α = 0.10 α = 0.05 α = 0.01
Large effect 20 26 38
(d = 0.8)
Medium effect 50 64 95
(d = 0.5)
Small effect 310 393 586
(d = 0.2)
Source: Cohen (1992), p. 158.

Recommendation: Use estimates of statistical power as a

guide to planning experiments (a priori power analysis).
Advanced Statistics
Statistical hypothesis testing
Power analysis

Is lack of statistical power a widespread

problem?
”We estimated the statistical power of the first and last
statistical test presented in 697 papers from 10 behavioral
journals ... On average statistical power was 13-16% to detect
a small effect and 40-47% to detect a medium effect. This is
far lower than the general recommendation of a power of 80%.
By this criterion, only 2-3%, 13-21%, and 37-50% of the tests
examined had a requisite power to detect a small, medium, or
large effect, respectively.”
Jennions, M.D., and A.P. Moeller 2003. Behavioral Ecology
14, 438-455.
Advanced Statistics
Statistical hypothesis testing
Power analysis

Further readings
Cohen, J. 1992. A power primer. Psychological
Bulletin 112: 155-159.

Jennions, M.D., and A.P. Moeller 2003. A survey of

the statistical power of research in behavioral
ecology and animal behavior. Behavioral Ecology
14: 438-455.

Hoenig, J.M., and D.M. Heisey 2001. The abuse of

power: the pervasive fallacy of power calculations
for data analysis. American Statistician 55: 19-24.
Advanced Statistics
Why multivariate analysis?

Why multivariate analysis?

Male Female
Accept 35 20
Refuse entry 45 40
Total 80 60
I Example: 44% of male applicants are admitted

by a university, but only 33% of female

applicants.
I Does this mean there is unfair discrimination?

I University investigates and breaks down figures

for Engineering and English programmes.

Advanced Statistics
Why multivariate analysis?

Simpson’s Paradox
Engineering Male Female English Male Female
Accept 30 10 Accept 5 10
Refuse entry 30 10 Refuse entry 15 30
Total 60 20 Total 20 40
I No relationship between sex and acceptance for
either programme. So no evidence of
discrimination. Why?
I More females apply for the English programme,
but it is hard to get into. More males applied
to Engineering, which has a higher acceptance
rate than English. Must look deeper than
single cross-tab to find this out!