Assignment 1

Homework Review

Assignment 1

Headlines / Key Takeaways

https://docs.google.com/document/d/1uSz_UfqQ1yPHI7R9I_x9qfn-VN-kiToLOkFDPq3eHO4/edit#

Anonymous

SurveyUSA Poll

Polling error: How one survey changed the Newsom recall campaign - CalMatters

SurveyUSA CEO Jay Leve admits that their earlier poll could have acted as an alarm clock. “It was not our intention to cause Democrats to suddenly sit bolt upright in their chairs,” he told CalMatters. And yet, “that, I think, happened.”

 

Mea Culpa - SurveyUSA

Key Takeaway

🔑 Know your data

Always spend some time with the primary source materials in addition to reporting around that source material.

In this case:

- Poll PDFs (Toplines / Crosstabs)

- Press Releases

- Conversations with experts

- Academic discussions & articles about those polls

- News about the polls themselves (like the CalMatters article)

✅ Looking for "apples-to-apples" comparisons

✅ Context on how often this has happened before

Relevant Example from FiveThirtyEight

✅ Explaining events around the key turning points, without making strong causal claims

⚠️ Reporting levels of precision in data

⚠️ Claims about the data's predictive value

⚠️ Causal Claims

⚠️ Keeping in mind the temporal element

Guest Speaker

Nathaniel Rakich

Intro to Stats for Journalists

source: http://www.differencebetween.net/language/words-language/difference-between-descriptive-and-inferential-statistics/

source: https://differencebetweenz.com/difference-between-descriptive-and-inferential-statistics/

source: https://www.geeksforgeeks.org/difference-between-descriptive-and-inferential-statistics/

Summary Stats

Mean

Examples

• https://fivethirtyeight.com/features/the-polls-are-all-right/

Ben Orlin — Math with bad drawings

Ben Orlin — Math with bad drawings

Weighted Average

Text

https://www.cuemath.com/weighted-average-formula/

Example

https://projects.fivethirtyeight.com/california-recall-polls/

Median

Examples

• Median Income
  https://www.nytimes.com/2014/04/23/upshot/the-american-middle-class-is-no-longer-the-worlds-richest.html
   
• Home Prices
  https://www.washingtonpost.com/business/2020/12/02/report-median-sales-price-houses-dc-now-exceeds-1-million/
   
• Mean vs Median - https://fivethirtyeight.com/features/americans-overestimate-the-number-of-muslims-and-teen-moms-but-not-by-as-much-as-reported/

Ben Orlin — Math with bad drawings

Ben Orlin — Math with bad drawings

Mode

Examples

• https://fivethirtyeight.com/features/stephen-colbert-most-frequent-guests-the-colbert-report/

Ben Orlin — Math with bad drawings

Ben Orlin — Math with bad drawings

Range

Ben Orlin — Math with bad drawings

Ben Orlin — Math with bad drawings

Correlation

Correlation

Pearson's Correlation Coefficient

Correlation doesn't imply causation, but it does waggle its eyebrows suggestively and gesture furtively while mouthing 'look over there'.

https://m.xkcd.com/552/

Examples

• https://fivethirtyeight.com/features/how-qi-and-za-changed-scrabble/
• https://fivethirtyeight.com/features/you-cant-trust-what-you-read-about-nutrition/

Ben Orlin — Math with bad drawings

Ben Orlin — Math with bad drawings

https://xkcd.com/925/

https://tylervigen.com/spurious-correlations

https://fivethirtyeight.com/features/you-cant-trust-what-you-read-about-nutrition/

Correlation & Causation

Standard Deviation

Variance

Examples

https://fivethirtyeight.com/features/election-update-the-polls-disagree-and-thats-ok/

Ben Orlin — Math with bad drawings

Ben Orlin — Math with bad drawings

Mystery Data

https://docs.google.com/spreadsheets/d/1fywjh2itH6p8Vpld8LAprdwXzsDByEhXiN5HqTQtpOM/edit#gid=0

What can you tell about these 4 mystery datasets with summary statistics?

Calculate:

- Mean

- Median

- Mode

- Correlation

- Variance

Distributions

https://fivethirtyeight.com/features/al-gores-new-movie-exposes-the-big-flaw-in-online-movie-ratings/

Normal Distribution

https://www.mathsisfun.com/data/standard-normal-distribution.html

Example

https://fivethirtyeight.com/features/all-the-cable-news-networks-are-covering-the-russia-story-just-not-the-same-one/

Plotting

Exploratory Data Visualization

Pearson correlation is ????

Pearson correlation is 0.9909

Editorial Choices

Let's Try It

https://docs.google.com/spreadsheets/d/1yKpTZCq9PbrwRzsPWjVT0qA2FM3n2s7QHM_z3EI6aSU/edit#gid=1323361134

• Which pollster is most / least accurate?
• How accurate are polls from:
  • Brown University
  • American Research Group
• Who is the most Republican-leaning pollster?

• Who is the most Democratic-leaning pollster?

• Are state polls less accurate than national polls?

• Which states has the most inaccurate presidential polling?

• How predictive are recall elections polls?

   

   

Let's Try It

https://docs.google.com/spreadsheets/d/1yKpTZCq9PbrwRzsPWjVT0qA2FM3n2s7QHM_z3EI6aSU/edit#gid=1323361134

• Which pollster is most / least accurate?
• How accurate are polls from:
  • Brown University
  • American Research Group
• Who is the most Republican-leaning pollster?

• Who is the most Democratic-leaning pollster?

• Are state polls less accurate than national polls?

• Which states has the most inaccurate presidential polling?

• How predictive are recall elections polls?

   

   

Let's Try It

https://docs.google.com/spreadsheets/d/1yKpTZCq9PbrwRzsPWjVT0qA2FM3n2s7QHM_z3EI6aSU/edit#gid=1323361134

• Which pollster is most / least accurate?
• How accurate are polls from:
  • Brown University
  • American Research Group
• Who is the most Republican-leaning pollster?

• Who is the most Democratic-leaning pollster?

• Are state polls less accurate than national polls?

• Which states has the most inaccurate presidential polling?

• How predictive are recall elections polls?

   

   

Let's Try It

https://docs.google.com/spreadsheets/d/1yKpTZCq9PbrwRzsPWjVT0qA2FM3n2s7QHM_z3EI6aSU/edit#gid=1323361134

• Which pollster is most / least accurate?
• How accurate are polls from:
  • Brown University
  • American Research Group
• Who is the most Republican-leaning pollster?

• Who is the most Democratic-leaning pollster?

• Are state polls less accurate than national polls?

• Which states has the most inaccurate presidential polling?

• How predictive are recall elections polls?

  
  

  
  

What makes a good pitch?

Dataset?      or a

I have a dataset I'm interested in

Question?

I have a journalistic question that I'm interested in trying to answer

Does your pitch start with a:

What questions will I ask?

Where will I get the data?

Quick Brainstorm

project group "speed dating" 

https://forms.gle/bzjmNvmXaqAmygjU7

• Headline of Pitch Brainstorm
  - Trash in New York City
   
• Questions
  - Where does NYC export it's trash to?
  - Does that differ by borough?
  - Are there private / public actors?
   
• Possible Datasets
  - NYC Trash data?
  - Google Maps?

Project Brainstorm

https://forms.gle/dJue4zuAFmewcwdw6