Data literacy: What, Why and How? Data and Statistics: the sciences, the literacies and collaborations - presented at the UN World Data Forum 2017

1. Data literacy: What, Why
and How?
Data and Statistics: the sciences,
the literacies and collaborations
Helen MacGillivray
President-elect, International Statistical Institute
Principal Fellow, Higher Education Academy
Australian Senior Learning and Teaching Fellow

2. Lessons from statistics
statistical literacy and statistical sciences
 What can be learnt for data literacy from decades of
promoting and efforts to enable statistical literacy?
 What can be learnt for data science from experiences
with statistical sciences?
 Experiences from learning and teaching
 Education across all levels
 Primary, junior secondary, upper secondary
 Tertiary: other disciplines and training of future professionals
 Workplace, professionals, researchers, managers, consultants
 Adults, society
 Education is both the challenge and the key
2

3. Lessons from statistics
- statistical literacy and statistical sciences
Some initial advice
 Descriptions can be constructive but definitions are not
 Discussion essential and enlightening but diagrammatic
representations are not
 Definitions and diagrammatic representations tend to
take time and attention aware from productive effort and
impose misleading and unnecessary boundaries
 (For example, anything involving Venn diagrams a waste of space)
3

4. Some descriptions of statistical literacy
 Statistical literacy is necessary for citizens to understand
material presented in publications such as newspapers,
television and the internet.
 Good “statistical citizens”: able to consume the information
that they are inundated with on a daily basis, think critically
about it, and make good decisions based on that
information. Some researchers call this “statistical literacy.”
Rumsey (2002)
 People’s ability to interpret and critically evaluate statistical
information and data-based arguments appearing in diverse
media channels, and their ability to discuss their opinions
regarding such statistical information (Gal 2000)
4

5. Some descriptions of statistical literacy
 A simple example from school levels
 Upper primary: read the data
 Junior secondary: read between the data
 Middle/senior secondary: read beyond the data
 The University of …… statistical literacy programme
 Module 1. Producing data
 Module 2. Describing, Clarifying and Presenting Data
 Module 3. Interpreting data
 Completing these modules will help you to develop the skills you
need to:
 look behind the data with which you are presented at University and
in your everyday experiences,
 ask why these data are being presented in those forms,
 ask what questions can be answered or what arguments are being
made with these data.
 As you work through these modules you should become much
more critical about the way data is produced, the way data is
presented and the way data is interpreted. 5

6. Some descriptions of data literacy
 Data literacy is the ability to read, create and
communicate data as information and has been
formally described in varying ways.
 The desire and ability to constructively engage in
society through and about data
http://datapopalliance.org/item/what-is-data-literacy/
 http://databrarians.org/2015/02/what-is-data-literacy/
 In Libraryland, “data literacy” seems to consist of two
aspects: information literacy and data management.
 Data literacy is the ability to interpret, evaluate, and
communicate statistical information…how statistical
information is created, encompassing data production
 Data management …. belongs to the data production
phase … perhaps one aspect of data literacy that can
be reserved for the specialists.
6

7. The what
 Descriptions constructive, definitions not
 Discussion of inter-relationships constructive because
different descriptions help everyone understand what and
why
 Any subset attempts or representations misleading and
waste of time.
7
The why
Descriptions give reasons:
for everyone to extent appropriate for level of education,
for training and work context

8. The how
RSS Centre for
Statistical Education
problem-solving cycle
Plan
Collect data
Process data
Discuss
Popularised by Wild and
Pfannkuch (1999). From
quality/industrial statistics
MacKay and Oldfield
(1994), Shewhart and
Deming (1986)
Based on data-handling
cycle UK National School
Curriculum 1970’s
(Holmes, 1997). Problem-
solving cycle (Marriott,
Davies and Gibson, 2009)
8
Statistical investigation process (under various terms, descriptions)
heart of statistical education and practice of statistics

9. From statisticians
Cameron (2009) considers
 desirable key components of university-based training
 consults what many “wide and experienced”
statisticians have written (e.g. Box, 1976)
 builds on Chambers’ (1993) “greater statistics”
 identifies
 formulating a problem so that it can be tackled statistically
 preparing data (including planning, collecting, organising and
validating)
 analysing data
 presenting information from data
 researching the interplay of observation, experiment and
theory.
 comments that such training is an appropriate
foundation for most statisticians wherever they may be
employed.
9

10. From statisticians
Kenett & Thyregod (2005)
 describe the 5 steps in statistical consulting
 problem elicitation
 data collection and/or aggregation
 data analysis using statistical methods
 formulation of findings & consequences
 presentation of findings and conclusions/recommendations.
 “important to take part in collection of data, or at least have
the opportunity to watch data being collected or generated.”
 and that not being involved in collecting data
 “has led some graduates to be of the opinion that taking part
in the collection of data is a waste of the statistician’s
precious time...implies risk of getting dirt on your hands.”
 “Our long-term objective is to encourage academic courses
to cover the full 1–5 cycle....especially steps 1, 2 and 5”
10

11. Statistical investigating at the heart of the
discipline, science and profession of statistics
Barnett (1986)
We see, tied up together, the role of the statistician
as consultant, consultancy as the stimulus for
research in statistics, and consultancy as the basis
for teaching statistics
Bisgaard and Bisgaard (2005) outline 3 roles: a pair of
hands; the expert; the catalyst/collaborator/coach
Cameron (2009): ‘entwined collaboration’ and ‘serial
collaboration’
The practice of statistics – from statisticians
11
Note advocacy for no division at
introductory tertiary; same foundation.
For example, Wild (2006), MacGillivray
(1998, 2005a), Cameron (2009)

12. The how
Some great work internationally, nationally and
locally, but
Penetration not great and problems persist. Why?
• Nature, pervasiveness of statistics; universality of
educational needs
• Dynamic nature of statistics: responds to data, technologies,
disciplines, workplaces
• Far too much of new ways of teaching old; other disciplines
• Technology resources
• Not enough real, complex, many-variable datasets
• Cobb (2015): Mere renovation is too little too late: we need to
rethink our undergraduate curriculum from the ground up
Penetration and implementation of these decades
of advocacy for statistical education for all and
for professional training?
12

13. Understanding the what of statistics
Some impediments: belief in other disciplines that can add
to basic background; calculations; trickle down effects
from research
Some impediments for data literacy and data science?
Similar – for ‘calculations’ substitute ‘coding’
• Statistics is the science of data, variation and
uncertainty.
• Statistics sources, evaluates, appraises, interrogates,
investigates, models, critiques, develops, applies,
interprets and communicates data, variation and the
information therein.
• Statistics works with, within and across all disciplines,
government, business, industry and society.
• Statistics and statistical thinking are pervasive,
universal and central to all evidence-based progress
and furthering of knowledge.
13

14. The how and collaboration
 Professionals need to get involved in the nitty gritty
 Observe, listen, communicate
 Enable coherent development over school
 Authentic working with other disciplines
 ASSESSMENT is key
 Simple to complex
 Real contexts, real data, complex data
 Technology resources for learning and assessment
 Look for cross-overs not boundaries
 Collaboration & sharing
14
Thank you & here’s to
statistics and data

15. Strengthening the ‘roots’ in undergraduate
curricula for future statisticians
 Heed long-time advocacy of professional statisticians
 Barnett (1986)
 “we see, tied up together, the role of the statistician as
consultant, consultancy as the stimulus for research in
statistics, and consultancy as the basis for teaching
statistics”.
 Authentic experience of full statistical investigation process:
 Cameron (2009) builds on Chambers’ (1993) ‘greater statistics’
 Kenett & Thyregod (2005) also describe 5 similar steps in
statistical practice/consulting
 “important to take part in collection of data, or at least have the
opportunity to watch data being collected or generated.”
 “encourage academic courses to cover the full 1–5
cycle....especially steps 1, 2 and 5.”
 Real, large contexts and data: simple within complex
 Maths as servant of statistics
 Technological and data systems know-how
15

16. Strengthening roots in school curricula
 Analyse what’s gone wrong in statistical education
‘reform’ over 2-3 decades
 New dogma for old
 New ways of learning old content & old sequencing e.g.
inference for means before proportions
 Domination of 1 and 2 variables and measures
 Toy datasets
 Lack of coherent development
 Domination of psychology thinking e.g. analysing
understanding of sampling distributions
 ‘The’ question & ‘the’ answer
 Need
 Variables, variation, visualisation
 Coherent development built up around types of variables
 Authentic full statistical data investigations – from simplest
 Simple within complex
16Thank you and here’s to statistics!