Evidence, Visualization, and Biases

Global Strategic Trends, part 1

Robert W. Walker

2025-11-19

A Prompt for the Day

KGW

What is the value of this?

Introducing causality….

Augmented Analysis with Claude

A Chat

The Plot in Esquisse

The Plot in Esquisse

Some Probability

Some Language

Sensitivity refers to the ability of a test to designate an individual with a disease as positive. Specificity refers to the ability of a test to designate an individual without a disease as negative.

False positives are then the complement/opposite of specificity and false negatives are the complement/opposite of sensitivity.

Truth Positive Test Negative Test
Positive Sensitivity False Negative
Negative False Positive Specificity

Examples:

  • Drug testing
  • Diagnostic testing

Applied to Hypothesis Testing

When we get to hypothesis testing in a few weeks, this comes up again with null and alternative hypotheses and the related decision.

Truth Reject Null Accept Null
Alternative Correct Type II error
Null Type I error Correct

A COVID test

Suppose a COVID test is 0.99 sensitive and 0.95 specific. At the time of administration, it is thought that roughly twelve percent of the population is infected.

What proportion of tests are positive?

  • Two positive tests:
    • True positives 0.99*0.12=0.1188 and
    • False positives 0.05*0.880=.044;
    • sum 0.1628.

What is the probability that a person with a positive test is infected?

  • Of .1628 positive tests, 0.1188 come from the infected:

\frac{0.1188}{0.1628} = 0.7297

Two Further Issues in Probability

  • Juries and Bayes Rule
  • Counting Rules

Counting Rules

  • Simple counting rules, k questions [True/False, n=2]: N^k
  • Multi-step experiments: Whataburger
  • Factorial designs
  • Combinations and permutations

Whataburger

Texas Fast Food Chain says there are 36,864 to order a Whataburger. How? - 2 burger patties (Regular, Junior) up to triple meat (R,J,RJ,RR,JJ,RRR,JJJ,RJJ,JRR) - 4 bread options - 3 condiments (mayonnaise, mustard, ketchup) - Vegetables (lettuce, tomato, pickle, onion) - Cheese, jalapeno, bacon 9*4*(1+3+3+1)*(1+4+6+4+1)*(1+3+3+1)

Just like the t-shirts say, 36,864 ways to make a whataburger!

citation

Meetings

Six students wish to meet with a professor… - Alice, Bob, Cat, Dharma, Ernest, Fred How many arrangements of the meetings into six slots? - Six ways for the first slot (ABCDEF) - Five ways for the second (whomever is not first) - Four ways for the third, and so on….

6*5*4*3*2*1=720

Permuations and Combinations

The key difference is whether or not the order matters. Permutations deem order as relevant whereas combinations do not. There are (at least as many or) more permutations than combinations.

  • Is the Powerball lottery a permutation or a combination?
  • What about a seating chart for this class?

Powerball

  • 59 white balls numbered 1 through 59
  • 35 red balls numbered 1 through 35
  • Choose 5 white balls and one red ball.

Calculating (old) Powerball

  • Order doesn’t matter for the chooser, you choose six numbers (five white, one red)
  • Then just apply the formula for combinations (but we need to account for the 35 red balls).

_nC_x = \frac{n!}{x!(n-x)!} = \frac{59!}{5!54!}

So the chances of winning the lottery are just a combinations problem with a second-step.

factorial(59)/(factorial(5)*factorial(54))*35
[1] 175223510

Ground Truth

Image

Image

New Powerball

  • Five white balls: 1 through 69
  • One red powerball: 1 through 26

Link to page

Random Variables

I do not love the book definition of this. Technically, it is a variable whose values are generated according to some random process; your book implies that these are limited to quantities.

It is really a measurable function defined on a probability space that maps from the sample space [the set of possible outcomes] to the real numbers.

A Core Idea: Independence

What does it mean to say something is independent of something else?

  • The simplest way to think about it is, “do I learn something more about x by knowing y than not”. If two things are independent, I don’t need to care about y if x is my objective.

General Ideas

  • Expectation

Expectation

Expectation
  • Variance

Variance

Variance

Part of 3.4 are incomplete.

  • Covariance
  • Linear Functions of Random Variables

Two Examples of Simpson’s Paradox for Homework

  • Discrimination
  • Admissions

Introducing Esquisse

Esquisse

Esquisse

Executing

esquisse:::esquisser(viewer="browser")

NB: It needs to run in a separate browser window.

Selecting Data

What is available in the environment?

A Run Through

Setting the Context for GST 7

Note the data of dissemination.

GST 6

Image

Image

Regions

MapESA

MapESA

MapSEA

MapSEA

MapCA

MapCA

MapSWA

MapSWA

MapAFR

MapAFR

MapRUS

MapRUS

PopEUR

PopEUR

MapCAC

MapCAC

MapNA

MapNA

Three Areas

  • Demographic Breakdown
  • Digital Usage
  • Climate, or ….

Demographics

PopESA

PopESA

PopSEA

PopSEA

PopCA

PopCA

PopSWA

PopSWA

PopAFR

PopAFR

PopRUS

PopRUS

PopEUR

PopEUR

PopCAC

PopCAC

PopNA

PopNA

Digital

DigESA

DigESA

DigSEA

DigSEA

DigCA

DigCA

DigSWA

DigSWA

DigAFR

DigAFR

DigRUS

DigRUS

DigEUR

DigEUR

DigCAC

DigCAC

DigNA

DigNA

Third Thing

ClimateESA

ClimateESA

ClimateSEA

ClimateSEA

ClimateCA

ClimateCA

ClimateSWA

ClimateSWA

EnerAFR

EnerAFR

WheatRUS

WheatRUS

MigEUR

MigEUR

ClimCAC

ClimCAC

DrugsNA

DrugsNA

The Arctic

Claims

Claims

Oil/Gas

Oil/Gas

Ships

Ships

Antarctica

Claims

Claims

Tourism

Tourism

The Oceans

Levels

Levels

Oil/Gas

Oil/Gas

Permits

Permits

Space/Orbits

Orbits

Orbits

Companies

Companies

Satellites

Satellites

Cyberspace

Usage

Usage

Information

Information

Incidents

Incidents

Themes

Society

Pop Dist

Pop Dist

Urbans

Urbans

World Religions

World Religions

Economics

GDP/MC

GDP/MC

Employment

Employment

Environmental Concerns

Global Temps

Global Temps

Energy

Energy

Info Tech

Global Transport

Global Transport

Causes

Causes

Conflict and Security

Regional Spend

Regional Spend

Minerals

Minerals