Of course techniques matter. Descriptive statistics and tests of significance in relation to sample data are crucial. Visualization - another point of agreement with Rachel - is enormously helpful. I even have some sympathy with straightforward use of regression techniques although always with more of an emphasis on strength or relationships than form. And multi-level modelling at least recognizes the causal powers of layered social systems.
My problem is with Q Step conceived as a programme which makes Sociology like Economics and UK Sociology like US Sociology and Political Science - dominated by quantitative pieces dealing with the banal or fitting models to complex social processes which are incapable of dealing with radical changes in context. I do not give a monkey's about UK Sociologists getting articles into US journals - a real triumph of formal over substantive rationality but important it seems to University managers (but then I am a well pensioned Emeritus so why would I).
In dealing with complex systems and changes in them, descriptive methods - data series understood as quantitative narratives - are the most useful - something exactly in the spirit of what Rachel is teaching. Then use classification techniques across time - again not hard to do and for me use Cluster Analysis rather than Latent Class Analysis with its reification of the causal powers of disembodied variables. Easy to do and interpretation depends entirely on contextual knowledge - mixed methods indeed. This can be done at all levels micro, meso and macro.
I have a cynical view which derives in part from my membership of the former ESRC Training Board and a sense of guilt in not having made enough fuss about severe reservations when I was actually on it. The combination of arrogance and ignorance of some of the proponents of 'complicated (not complex I tell you) methods will solve our problems' infuriated me then and does so more now.
All the above is academic politics - arguments about what we should be teaching our students. But there is a real politics as well. Economics hid its nonsense behind what Joan Robinson described as a thicket of algebra for years and the forms it developed were and still are used as an intellectual justification for neo-liberalism's marketization of everything. This was not just a matter of the deductive nonsense of positive economics after Samuelson et al. It also affected the ways in which economic measurements - econometrics - were used in modelling in relation to decisions. Numbers matter a great deal. I am all for a Q Step conceived on Rachel's lines but I don't think that was what Nuffield and the ESRC has in mind although subversion is happening and should happen more BUT NOT EVERYWHERE OR EVEN IN MOST PLACES.
David Byrne
From: Rachel Cohen [[log in to unmask]]
Sent: 19 March 2015 08:54
To: BYRNE D.S.; [log in to unmask]
Subject: Re: Q step
Sent: 19 March 2015 08:54
To: BYRNE D.S.; [log in to unmask]
Subject: Re: Q step
I feel like I should probably respond to this, because:
a) I am a q-step centre coordinator
b) I am involved in radstats - as one of the journal editors
b) I consider myself a mixed-methods researcher
So briefly on the logic and effects of q-step. First off, I would agree that some of the rush for quants can be simplistic. The focus on 'business needs' contrary to the logic of many in radstats (or the idea of education as developing the citizen etc). There is also a frustrating focus on 'excellence' which I find as problematic when related to student development as in its other guises. But, I started getting involved in teaching quantitative methods because I do believe in the importance of numeracy and find the ease with which academic colleagues and educated friends brush off numeric information as being 'too complex', 'not for me' or 'all rubbish' frustrating.
So, how has q-step worked in practice (and I would note that some of this will be specific to my context, a non-Russell Group university, City University London)? First the 'data literacy' branch of q-step, which has been our biggest focus has provided an impetus to transform the undergraduate curriculum and introduce a focus on the student as reseracher and making links between theory and practice. As such we have introduced both quantitative, but also qualitative teaching from Year 1. The quants modules are called Lies Damned Lies and Statistics: Making sense of quantitative data (and includes lots of critique as well as more traditional questions about analysis) and the other is Producing Social Data and involves both questions about where data come from (political and social questions) as well as 'how to' aspects of data production. Neither includes much formal maths. Other new modules will include, at year 2, data visualisation taught by a colleague from Infomatics who's very excited about about embedding informatics in social science(!), and various at Year 3 (but these are two years away so I can't provide much detail!). As in all Q-step centres there's also been an effort to embed quantitative (and we've emphasised, qualitative) methods - or practical thinking about data - into other substantive and theoretical modules. Second, with respect to the focus of q-step on developing pathways for students to do more advanced quantitative methods, at City we specifically chose not to recruit students who have A-Level maths, but rather to assume that our current student intake (mostly BME, female, low SES, Londoners) are capable of understanding and producing quantitative analysis. Since we know that these students have arrived with an interest in sociology, criminology etc, and have taken modules in both quant and qual methods in year 1, by giving them the option to move to a quantitative pathway at year 2, we're hopefully focusing on people with substantive research interests, not simply a desire to produce models. This may or may not work. It's still early stages.
None of this resolves all the issues raised by David or by Brian, in his longer article, but hopefully it highlights that the implementation of q-step may not be about maths (definitely not about formal maths), may simultaneously improve qualitative methods (even mixed methods, although at present this is included relatively informally at City), may be critical bout the production and use of numbers, may engage with new (non linear) methods, and may encourage previously socially disadvantaged students who enter university with critical social quesions to develop and make sense of their own research.
So there are lots of problems, but I would also say that there are many aspects of the Q-Step programme that can be used to trigger a more creative, hands on, critical curriculum. And while it won't address complexity in all its complexity(!), in a three year degree with students taking classes across a wide range of topics, I feel like it probably does more to help than hinder.
Best, Rachel.
a) I am a q-step centre coordinator
b) I am involved in radstats - as one of the journal editors
b) I consider myself a mixed-methods researcher
So briefly on the logic and effects of q-step. First off, I would agree that some of the rush for quants can be simplistic. The focus on 'business needs' contrary to the logic of many in radstats (or the idea of education as developing the citizen etc). There is also a frustrating focus on 'excellence' which I find as problematic when related to student development as in its other guises. But, I started getting involved in teaching quantitative methods because I do believe in the importance of numeracy and find the ease with which academic colleagues and educated friends brush off numeric information as being 'too complex', 'not for me' or 'all rubbish' frustrating.
So, how has q-step worked in practice (and I would note that some of this will be specific to my context, a non-Russell Group university, City University London)? First the 'data literacy' branch of q-step, which has been our biggest focus has provided an impetus to transform the undergraduate curriculum and introduce a focus on the student as reseracher and making links between theory and practice. As such we have introduced both quantitative, but also qualitative teaching from Year 1. The quants modules are called Lies Damned Lies and Statistics: Making sense of quantitative data (and includes lots of critique as well as more traditional questions about analysis) and the other is Producing Social Data and involves both questions about where data come from (political and social questions) as well as 'how to' aspects of data production. Neither includes much formal maths. Other new modules will include, at year 2, data visualisation taught by a colleague from Infomatics who's very excited about about embedding informatics in social science(!), and various at Year 3 (but these are two years away so I can't provide much detail!). As in all Q-step centres there's also been an effort to embed quantitative (and we've emphasised, qualitative) methods - or practical thinking about data - into other substantive and theoretical modules. Second, with respect to the focus of q-step on developing pathways for students to do more advanced quantitative methods, at City we specifically chose not to recruit students who have A-Level maths, but rather to assume that our current student intake (mostly BME, female, low SES, Londoners) are capable of understanding and producing quantitative analysis. Since we know that these students have arrived with an interest in sociology, criminology etc, and have taken modules in both quant and qual methods in year 1, by giving them the option to move to a quantitative pathway at year 2, we're hopefully focusing on people with substantive research interests, not simply a desire to produce models. This may or may not work. It's still early stages.
None of this resolves all the issues raised by David or by Brian, in his longer article, but hopefully it highlights that the implementation of q-step may not be about maths (definitely not about formal maths), may simultaneously improve qualitative methods (even mixed methods, although at present this is included relatively informally at City), may be critical bout the production and use of numbers, may engage with new (non linear) methods, and may encourage previously socially disadvantaged students who enter university with critical social quesions to develop and make sense of their own research.
So there are lots of problems, but I would also say that there are many aspects of the Q-Step programme that can be used to trigger a more creative, hands on, critical curriculum. And while it won't address complexity in all its complexity(!), in a three year degree with students taking classes across a wide range of topics, I feel like it probably does more to help than hinder.
Best, Rachel.
On 18/03/2015 16:25, BYRNE D.S. wrote:
****************************************************** 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. *******************************************************
Jeff Evans mentioned Q Step in his recent email re possible books / publications from Radstats. I have serious issues with the rationale behind Q Step although I accept that some local centres are doing decent things. For a context on this I recommend Brian Castellani's piece in Society Today - see:
http://www.discoversociety.org/2014/11/04/focus-complexity-and-the-failure-of-quantitative-social-science/
Bluntly put the conventional quantitative programme across the social sciences has, as the Gulbenkian Commission on the Future of the Social Sciences noted in 1996, very largely failed. This is not an attack on statistics as numbers - vital for social description, nor on the sensible and limited use of exploratory approaches but on the dominance of conventional linear methods and the belief that the social sciences have to be capable of mathematical formalism to be proper sciences. So Q Step comes along and says we solve this by training undergraduates in more formal mathematics and 'advanced' statistical methods. The cruder exponents of this position, not I think to be found among the Radstats membership, refer to the great success of the quantitative programme in Economics to encourage this trend!!
On the whole Radstats has not had a lot to say about methods although there was some decent discussion in Demystifying Social Statistics. I am struck by how there is not so much a debate about these issues as an absence of one. For example I think that a lot of what is done at the Cathie Marsh Centre in Manchester is done ignoring the points Marsh herself made in her wonderful book on Surveys in Social Research.
We would all agree that innumeracy is not a valid epistemological position otherwise we would not be in Radstats but there is a politics of methods as well as a politics of numbers. The politics of methods is first an internal politics of the Academy but it has implications which go beyond it. I think the politics of numbers matters a lot more since most of the numbers are based on nothing more complicated than counting and that is crucial but the politics of methods matters as well.
David Byrne
|
This email has been checked for viruses by Avast antivirus software. |