Dear Jane,

If by 'non-specialist reader', you mean any literate person, I'm not sure there is a way to make quantitative data universally intelligible without some effort on the reader's part.  However, there are some well known guidelines, for example:
- using a histogram or bar chart, rather than a line graph, to display data on categorical variables
- line graphs are appropriate for regular time series, but where observations are irregular, may give the wrong impression and a bar chart may be more appropriate
- ensuring that ticks are equally spaced and adequately labelled
- when dividing a population according to a single variable, I don't see any big problems with a carefully labelled pie chart

I think the reason we find graphical representations of data in popular publications is that the authors believe that the audience can't relate to tabular data, and probably often enough, the authors themselves can't either.  

What concerns me more is the verbal presentation of data.  For example, I've read that ' about a third of adults are without work'.  In reality, something like a third of adults were unemployed or underemployed, but of course the underemployed were not without work and due to the low participation rate in the population under discussion, about 3/4 of adults were actually 'without work'.  Yesterday I read that the Turkish Statistical Institute deliberately undercounts the unemployed because they only include those looking for work and exclude discouraged jobseekers!  Just about every month you read that 'x jobs were lost last month', when in reality, x fewer persons were employed last month than in the preceeding month.  

A couple of years ago, there was a report in Metro 'Almost half the black children in Britain are being raised by single parents, new Government figures reveal.'  Actually, almost half of Black FAMILIES or one parent families - two parent families tend to have more children, according to the same source, presumably the ONS's 'Social trends 2007'.

That kind of thing.  

What I try to do is to explain what statistics actually aim to collect, why they do it the way they do (e.g. excluding discouraged jobseekers), and how we can use official statistics for our own analysis.

In solidarity,
Harry

-----Original Message-----
From: email list for Radical Statistics [mailto:[log in to unmask]] On Behalf Of Jane Galbraith
Sent: Thursday, 19 March 2009 21:18
To: [log in to unmask]
Subject: Re: WHY MORE EQUAL SOCIETIES ALMOST ALWAYS DO BETTER

Two important problems have arisen in this correspondence:
1) How can we improve the visual presentation of information for the
non-specialist reader? (Some discussion below)
2) How should we balance scepticism about the quality of survey (and
other) data but still encourage its proper use in understanding the world
and formulating policy? (Discussion not included below)

An argument may use a diagram as an illustration; this would be schematic
rather than a representation of data so axis labels such as LOW or HIGH
are adequate.

A graph, however, on which data is plotted should satisfy some basic
requirements. The least of these is that the variables plotted should be
clearly identified. The graphs on the website

http://www.equalitytrust.org.uk/why/evidence/methods

do not (as Robert points out) say which measure of inequality is plotted.

The graph would not look more complicated if it had a caption saying what
exact measures were plotted and tick marks giving at least the range of
values on the horizontal axis. But keep the LOW HIGH labels too as a
visual aid to highlight the main message.
(A scatterplot matrix on the website would be far more use to anyone
wanting to know about the effects of inequality.)

Rather than throw up our hands in horror at the ignorance of the public we
(and the authors, Richard Wilkinson and Kate Pickett) should use every
opportunity to use simple but well designed graphs. The public need to be
exposed to good practice. People will make an effort to understand a graph
if they are interested in its content. By referring in the text to
features in a graph the authors can help the reader to study it. Bar
charts and time-series plots are widely used in the press. We need to
encourage the use of histograms and scatterplots.... (and discourage
pie-charts!).

Happy graphical presentation of information to you all

Jane

PS I hope that Martin Bland is passing on the Rad Stats correspondence to
Kate Pickett and that she will pass it on to her publishers!

-- 
Mrs Jane Galbraith
Honorary Research Associate
Department of Statistical Science
University College London
Gower Street
London WC1E 6BT

[log in to unmask]





> Kate Pickett's response made depressing reading - presumably the
inability
> to read a graph and understand statistics extends to those 'high
educated
> people' who run the economy?
>
> Did Putnam publish his data in peer reviewed journals? I found myself
getting very angry when reading 'Bowling Alone' because at crucial
junctures in the argument there were no hard data (i.e. actual numbers).
I
> nearly gave up reading it - but obviously it was too important to do
that.
> Not quite laymen perhaps, but I mentioned 'Bowling Alone' when doing a
seminar with a bunch of clergy - I was immediately challenged on
Putnam's
> data and asked questions about sample size, representativeness of
samples,
> confidence intervals etc. Some of them had read the book and looked in
the
> back for the data and found them wanting.
>
> ... but still, you can't please all of the people all  the time.
>
> Robert
>
>
>
>
> Professor Robert Moore
> School of Sociology and Social Policy
> Eleanor Rathbone Building
> The University of Liverpool
> L69 7ZA
>
> Telephone and fax: 44 (0) 1352 714456
> ________________________________________
> From: email list for Radical Statistics [[log in to unmask]] On
Behalf Of Bland, M. [[log in to unmask]]
> Sent: 17 March 2009 09:40
> To: [log in to unmask]
> Subject: Re: WHY MORE EQUAL SOCIETIES ALMOST ALWAYS DO BETTER
>
> Kate Pickett (an epidemiologist, and certainly not afraid of statistics)
is a colleague, with an office just down the hall from me. I forwarded
Ray's email to her. She has sent this reply:
>
> The reason for the lack of statistical detail in our book is because it
is a popular book, aimed at the layperson, and we found (somewhat to our
surprise) that a lot of highly educated people who saw the book in
manuscript didn't understand graphs, let alone statistics. We modelled
the final charts in the book on those used by Putnam in "Bowling Alone",
which people found easy to comprehend. For anybody interested in more
detail, all our methods and statistics are on The Equality Trust
website, and we gave a reference to this in the book.
>
> They are available at:
> http://www.equalitytrust.org.uk/why/evidence/methods
>
> In addition, almost all of our findings have been previously published
in peer-reviewed journals, again references to these are made in the
book, and some, where copyright permits, can also be downloaded from the
website.
>
> I'm not on the list, but would be happy for you to post this reply from
me. Far from being disdainful of statistics, we view them as a vital
social microscope which allows us to see patterns that are not apparent
to the naked eye. We live and breathe statistics, but sadly most people
don't, and we wanted to communicate our message to the broadest
> readership possible.
>
> Best, Kate.
>
> There you have it!
>
> Martin
>
> Ray Thomas wrote:
>> Why more equal societies almost always do better is the subtitle of a
gallant book by Richard Wilkinson and Kate Pickett.
>> In spite of the emphasis given by use of the word "equal" the authors
appear to be scornful of statistic. There are plenty of charts in the
book but most of them do not include statistics for the scales used.
Typically the scales are labelled 'low' at the bottom of the vertical
scale and high at the top end. Consistently the LH of the horizontal
scale is labelled 'low' and the Right hand of the horizontal scale is
label high. Typically the charts do not give the units of measurement.
Can anyone explain the apparent disdainful attitude to statistics that
seems to contradict the theme of the book?
>> Ray Thomas

******************************************************
Please note that if you press the 'Reply' button your
message will go only to the sender of this message.
If you want to reply to the whole list, use your mailer's
'Reply-to-All' button to send your message automatically
to [log in to unmask]
Disclaimer: The messages sent to this list are the views of the sender and cannot be assumed to be representative of the range of views held by subscribers to the Radical Statistics Group. To find out more about Radical Statistics and its aims and activities and read current and past issues of our newsletter you are invited to visit our web site www.radstats.org.uk.
*******************************************************

Notice:

The information contained in this email message and any attached files may be confidential information, and may also be the subject of legal professional privilege. If you are not the intended recipient any use, disclosure or copying of this email is unauthorised. If you received this email in error, please notify the DEEWR Service Desk by calling 1300 305 520 and delete all copies of this transmission together with any attachments.

******************************************************
Please note that if you press the 'Reply' button your
message will go only to the sender of this message.
If you want to reply to the whole list, use your mailer's
'Reply-to-All' button to send your message automatically
to [log in to unmask]
Disclaimer: The messages sent to this list are the views of the sender and cannot be assumed to be representative of the range of views held by subscribers to the Radical Statistics Group. To find out more about Radical Statistics and its aims and activities and read current and past issues of our newsletter you are invited to visit our web site www.radstats.org.uk.
*******************************************************