1 of 63

Bringing Data Science Communication into the Classroom

Lucy D’Agostino McGowan

Wake Forest University

@LucyStats

2 of 63

follow along

lucymcgowan.com/talk

3 of 63

Data from my students

Theory on statistical communication

How to + examples

1

2

3

4 of 63

89%

94%

83%

Would recommend keeping the communicating statistics section in future iterations of this course.

Found the three lectures on communicating statistics valuable.

Felt the communicating statistics lectures fit in well with the rest of the course.

5 of 63

6 of 63

How true it is

7 of 63

How true it is

TRUE

TRUE

8 of 63

How true it is

NOT TRUE

NOT TRUE

TRUE

TRUE

9 of 63

How interesting it is

NOT TRUE

NOT TRUE

TRUE

TRUE

10 of 63

How interesting it is

INTERESTING

INTERESTING

NOT TRUE

NOT TRUE

TRUE

TRUE

11 of 63

How interesting it is

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

NOT TRUE

NOT TRUE

TRUE

TRUE

12 of 63

Not interesting

Not true

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

13 of 63

Not interesting

Not true

Not really a problem

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

14 of 63

Not interesting . true

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

15 of 63

. interesting Not true

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

16 of 63

. interesting . true

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

17 of 63

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

18 of 63

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

19 of 63

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

20 of 63

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

21 of 63

Estimating COVID-19 mortality for patients on a mechanical ventilator

22 of 63

Estimating COVID-19 mortality for patients on a mechanical ventilator

Richardson S, Hirsch JS, Narasimhan M, et al. Presenting Characteristics, Comorbidities, and Outcomes Among 5700 Patients Hospitalized With COVID-19 in the New York City Area. JAMA. 2020;323(20):2052–2059. doi:10.1001/jama.2020.6775

23 of 63

Estimating COVID-19 mortality for patients on a mechanical ventilator

Clarification of Mortality Rate and Data in Abstract, Results, and Table 2. JAMA. 2020;323(20):2098. doi:10.1001/jama.2020.7681

24 of 63

The journey to true

MATHEMATICALLY

CORRECT

25 of 63

A tutorial to help folks recognize bad science in the news

26 of 63

A tutorial to help folks recognize bad science in the news

https://ed.ted.com/lessons/this-one-weird-trick-will-help-you-spot-clickbait-jeff-leek-and-lucy-mcgowan

27 of 63

A tutorial to help folks recognize bad science in the news

https://ed.ted.com/lessons/this-one-weird-trick-will-help-you-spot-clickbait-jeff-leek-and-lucy-mcgowan

“This one weird trick will help you spot clickbait”

28 of 63

A tutorial to help folks recognize bad science in the news

https://ed.ted.com/lessons/this-one-weird-trick-will-help-you-spot-clickbait-jeff-leek-and-lucy-mcgowan

“This one weird trick will help you spot clickbait”

29 of 63

The journey to true

MATHEMATICALLY

CORRECT

MARKETED CORRECTLY

30 of 63

Estimating COVID-19 mortality for patients on a mechanical ventilator

Richardson S, Hirsch JS, Narasimhan M, et al. Presenting Characteristics, Comorbidities, and Outcomes Among 5700 Patients Hospitalized With COVID-19 in the New York City Area. JAMA. 2020;323(20):2052–2059. doi:10.1001/jama.2020.6775

31 of 63

Estimating COVID-19 mortality for patients on a mechanical ventilator

Clarification of Mortality Rate and Data in Abstract, Results, and Table 2. JAMA. 2020;323(20):2098. doi:10.1001/jama.2020.7681

32 of 63

The journey to true

MATHEMATICALLY

CORRECT

DISSEMINATED CORRECTLY

MARKETED CORRECTLY

33 of 63

Examining COVID-19 cases per 100k in Georgia by county

34 of 63

Examining COVID-19 cases per 100k in Georgia by county

35 of 63

Examining COVID-19 cases per 100k in Georgia by county

36 of 63

The journey to true

MATHEMATICALLY

CORRECT

DISSEMINATED CORRECTLY

MARKETED CORRECTLY

AUDIENCE INTERPRETS CORRECTLY

37 of 63

New York Times Analysis

  • Extracted all articles from the New York Times Archive API from 2017-July 2021

  • Identified statistical terms in the lead paragraphs of each article

  • Summarized over month

38 of 63

39 of 63

40 of 63

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

41 of 63

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

42 of 63

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

43 of 63

TRUE

NOT TRUE

TRUE

NOT TRUE

INTERESTING

INTERESTING

NOT INTERESTING

NOT INTERESTING

44 of 63

A collaborator is trying to decide how to analyze observational data

45 of 63

A collaborator is trying to decide how to analyze observational data

46 of 63

A collaborator is trying to decide how to analyze observational data

OMG check out my dissertation!!

47 of 63

A collaborator is trying to decide how to analyze observational data

TRUE

NOT INTERESTING

48 of 63

A collaborator is trying to decide how to analyze observational data

TRUE

NOT INTERESTING

TRUE

INTERESTING

49 of 63

A collaborator is trying to decide how to analyze observational data

50 of 63

A collaborator is trying to decide how to analyze observational data

OMG check out my dissertation blog post!!

51 of 63

The journey to interesting

From a long pdf to a short blogpost

MEDIUM

52 of 63

The journey to interesting

From a long pdf to a short blogpost

MEDIUM

Only include the important stuff

LENGTH

53 of 63

The journey to interesting

From a long pdf to a short blogpost

MEDIUM

Only necessary equations, ditch the jargon

COMPLEXITY

Only include the important stuff

LENGTH

54 of 63

The journey to interesting

From a long pdf to a short blogpost

MEDIUM

Only necessary equations, ditch the jargon

COMPLEXITY

Only include the important stuff

LENGTH

Pick a single message

FOCUS

55 of 63

The journey to interesting

From a long pdf to a short blogpost

MEDIUM

Only necessary equations, ditch the jargon

COMPLEXITY

Humans like

pictures

VISUALS

Only include the important stuff

LENGTH

Pick a single message

FOCUS

56 of 63

The journey to interesting

From a long pdf to a short blogpost

MEDIUM

Only necessary equations, ditch the jargon

COMPLEXITY

Humans like

pictures

VISUALS

Only include the important stuff

LENGTH

Pick a single message

FOCUS

Promote it!

MARKETING

57 of 63

The journey to true

MATHEMATICALLY

CORRECT

DISSEMINATED CORRECTLY

MARKETED CORRECTLY

AUDIENCE INTERPRETS CORRECTLY

58 of 63

The journey to interesting and true

MEDIUM

COMPLEXITY

MATHEMATICALLY CORRECT

LENGTH

FOCUS

VISUALS

MARKETED CORRECTLY

DISSEMINATED CORRECTLY

AUDIENCE INTERPRETS CORRECTLY

59 of 63

Consider the Audience

MEDIUM

COMPLEXITY

MATHEMATICALLY CORRECT

LENGTH

FOCUS

VISUALS

MARKETED CORRECTLY

DISSEMINATED CORRECTLY

AUDIENCE INTERPRETS CORRECTLY

60 of 63

Consider the Content

MEDIUM

COMPLEXITY

MATHEMATICALLY CORRECT

LENGTH

FOCUS

VISUALS

MARKETED CORRECTLY

DISSEMINATED CORRECTLY

AUDIENCE INTERPRETS CORRECTLY

61 of 63

Consider the Marketing

MEDIUM

COMPLEXITY

MATHEMATICALLY CORRECT

LENGTH

FOCUS

VISUALS

MARKETED CORRECTLY

DISSEMINATED CORRECTLY

AUDIENCE INTERPRETS CORRECTLY

62 of 63

The journey to interesting and true

MEDIUM

COMPLEXITY

MATHEMATICALLY CORRECT

LENGTH

FOCUS

VISUALS

MARKETED CORRECTLY

DISSEMINATED CORRECTLY

AUDIENCE INTERPRETS CORRECTLY

63 of 63

Bringing data science communication into the classroom

Lucy D’Agostino McGowan

Wake Forest University

@LucyStats

  • Presentation template by Slidesgo
  • Icons by Flaticon
  • Infographics by Freepik
  • Images created by Freepik