Hi Folks there is a big statistical issue lying in here around Parttime work status.
If we 'include' this among the independent variables in a wage regression, it 'takes up' a lot of 'causal weight'.
But in reality it is endogenous. Working part-time is in turn affected by other matters which are among the independent variables.
In my work on the gender pay gap in UK I have omitted the variable, WORKS PARTTIME.
Overall I think that's best, because it is a conjuncture which is not independent of distal causes.
But when we omit 'Number of Children Living at Home', which in a wage regression is only an indirect (ie distal) cause,
(as often happens outside Economics when the statistician uses a Wage Regression among Workers but omits the 'Economically inactive',
and avoids using the Mills Ratio or Heckman selectivity formulation, through a cut in the reference population), we get reduced explanatory power.
We then have omitted a major proxy for women doing child care more than men: we have omitted both Children and Parttime.
So in some cases, when Children is not in the equation, we may need Parttime, but it is now a proxy for the presence of childcare responsibilities
and hence it represents the whole of society's stereotypical expectation, That a woman is more responsible than a man for these varied duties, including cooking, cleaning and so on.
Ironically the meaning of the coefficient on 'Parttime' is different for those with, and without, children at home.
I am worried that in Statistics textbooks the advice is shorn of all real meaning and reference. The variables are just seen as part of a correlation matrix, and
not in their real world social representation role, which is (here) a mixture of proxy, proximate, distal, endogenous, stereotypical, and so on which are all causal, but
they get all mixed up.
The variables are not just representing the thing which is in their name, e.g. WORKS PARTTIME. They also represent the causal mechanism that we are trying to estimate as the regression coefficient,
e.g. lost wages if the person works parttime.
If we don't train our statisticians both IN and OUTSIDE economics to be sensitive to al this, we may end up
with journal editors having ridiculous arguments about one thing (endogeneity) and ignoring the real issue ( social stereotyping and its real effects).
This is why I keep talking about being a 'realist'. The only way to choose the X variables for a statistical regression is to develop a working knowledge of all the proposed X factors, and then start to
choose or develop a representation.
FYI The Heckman formulation:
Eq 1. Wage = F ( X, IMR)
IMR is the inverse risk of being active in the labour market, the Mills Ratio.
Now call it lambda.
Lambda = G ( X, Z)
Z includes having children living at home.
Wendy Olsen
This message was sent from Wendy Olsen
Social Statistics Discipline
(and Cathie Marsh Institute)
Room G20 Humanities Building
University of Manchester
Manchester M13 9PL
Next big events: Sept. 12-13th we are in Panel 1, Labour as Method, at the DevtStudiesAssoc. in Oxford. our talk is at 2 pm in Magdalen College common room.
Facebook group on Integrated Mixed Methods Network, all welcome (it says 'closed group', just ask, and you can join) QCA TRAINING is ongoing.
________________________________________
From: email list for Radical Statistics [[log in to unmask]] on behalf of Paul Bivand [[log in to unmask]]
Sent: 27 September 2016 12:54
To: [log in to unmask]
Subject: Re: Gender pay gap = "Too few senior women"?
One of the major issues is that professional/managerial jobs are still very hostile to part-time working - and particularly for promotion/recruitment of women who want to work part-time. Public sector ones have been a bit of an exception to that to some extent only.
This leads to lots of graduate women working well below their skill level in jobs that do provide part-time opportunities fitting round families/caring responsibilities (like admin/clerical jobs or even retail). It's one of the main drivers in what gets called 'overqualification' though the researchers in the area seem strangely reluctant to even look at the gender/family situations.
Because women who want to work part-time are forced out of their career area, getting back in and securing promotion to a level equivalent to what they would have been without leaving is very difficult.
Paul Bivand | Associate Director of Analysis & Statistics | Learning and Work Institute
t. 020 7840 8335 | tw. @LWpaulbivand
e. [log in to unmask]
www.learningandwork.org.uk | @LearnWorkUK | 0116 204 4200
The contents of this e-mail and any attachment(s) are intended solely for the use of the addressee. If you are not the intended recipient please return the e-mail to the sender and delete from your mailbox.
-----Original Message-----
From: email list for Radical Statistics [mailto:[log in to unmask]] On Behalf Of jay ginn
Sent: 27 September 2016 08:59
To: [log in to unmask]
Subject: Re: Gender pay gap = "Too few senior women"?
I'm surprised at the suggestion that if there were equal proportions of men and women in senior jobs this would reduce or even remove the gender pay gap.
Tracey Warrens article ,A privileged pole, showed that even women with the advantage of a degree and no breaks for childcare earned less than similarly qualified men and this translated into smaller private pensions, Hope I've remembered correctly.
As others rightly say, caring breaks and periods in pt work contribute to women's lower lifetime earnings and my own research in ,degrees of freedom, 2002, shows how women's median earnings, ft emp rate and private pension membership all take a dive when there is child aged under 5 , at each educational level of women using Ghs for cross sectional synthetic life course. Joshi and Davies used longitudinal data to model hypothetical lifetime earnings for several typical female life courses, calculating losses due to motherhood. Though eldercare is also relevant.
Concerning pensions, Dave, reliance on derived pensions from a spouse is risky individually and unsatisfactory as policy, anachronistic. Divorced women are the poorest group of pensioners as most of them had children to raise alone. And a growing group.
While state pensions are catching up, belatedly, with changing family forms, private pensions still transmit women's typical life course into low or no such pension, on average 53% of meNs for those over 65 who have any, including widows. That was 2001 Data see article Social Trends 2004, but has it changed since then?
Jay
Sent from my iPad
> On 26 Sep 2016, at 10:22, John Bibby <[log in to unmask]> wrote:
>
> I've just been listening to Woman's Hour which suggested that much of today's gender pay gap can be explained by there being "Too few senior women". One spokeswoman even suggested that if you control for this, women often earn more.
>
> Is this storyline sustainable with existing statistics, does anyone know? If not, how much of the gender pay gap can be 'explained' by seniority etc.?
>
> I do not know what the best sources for this are. It used to be the "New Earnings Survey (NES)", which if it exists at all is now far from new.
>
> JOHN BIBBY
> ****************************************************** 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.
> *******************************************************
******************************************************
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.
*******************************************************
NIACE and the Centre for Economic and Social Inclusion have merged to become Learning and Work Institute. Please follow our Twitter account @LearnWorkUK for the latest news.
This message contains confidential information and is intended only for the individual(s) addressed in the message. If you are not the named addressee, you should not disseminate, distribute, or copy this e-mail. If you are not the intended recipient, you are notified that disclosing, distributing, or copying this e-mail is strictly prohibited.
******************************************************
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.
*******************************************************
******************************************************
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.
*******************************************************
|