{"id":978,"date":"2012-10-22T22:00:17","date_gmt":"2012-10-22T22:00:17","guid":{"rendered":"http:\/\/johncanning.net\/wordpress\/?p=978"},"modified":"2012-10-22T22:00:17","modified_gmt":"2012-10-22T22:00:17","slug":"978","status":"publish","type":"post","link":"https:\/\/johncanning.net\/wp\/?p=978","title":{"rendered":"Five (possible) barriers to quantitative methods in the social sciences"},"content":{"rendered":"

The news that the Nuffield foundation, the ESRC and HEFCE are to invest \u00a315.5m in centres to develop quantitative skills in the social sciences is very welcome. Nuffield foundation director Sharon Witherspoon\u2019s article, why the social sciences need a skills step-change<\/a> in the Guardian was published the same day that Society Counts<\/em> <\/a>was launched by the British Academy.<\/p>\n

My university degrees (BA, MSc, PhD) are all in geography, a discipline which underwent a \u2018quantitative revolution\u2019 in the 1960s. At the risk of over simplifying the history of geography the quantitative revolution can trace its routes to Fred K Schaefer\u2019s posthumously published 1953 article \u201cExceptionalism in Geography: A Methodological Examination in the Annals of the Association of American Geographers<\/a>\u201d. \u00a0Schaefer\u2019s sudden death prior to the article\u2019s publication meant that he was unable to respond to or defend criticism of his \u2018scientific approach\u2019 from the likes of Richard Hartshorne, yet there was no shortage of geographers willing and able to build on Schaeffer\u2019s idea. According to the more simplistic narratives of the history of geography, quantitative approaches were gradually edged out during the 1970s as more behaviourist and qualitative approaches took over.\u00a0 By the time I arrived at university as an undergraduate in 1994 statistics was very much, in Witherspoon\u2019s words, a \u2018bolt-on\u2019 module. (Please don\u2019t cite this blog post as an authoritative reference for the history of geography.)<\/p>\n

I\u2019ll leave the relative merits of different approaches \u00a0\u00a0to one side, but I\u2019ll share some thoughts about why quantitative approaches are frequently rejected.<\/p>\n

\"\"<\/a><\/p>\n

    \n
  1. There is an adage that if you can add, subtract, multiply and divide you can do statistics. If we are talking about the mechanics of undertaking statistical tests then there is a degree of truth in this, but how many beginners\u2019 statistics texts adequately explain the normal distribution, z scores or standard deviations? This stock diagram always appears in some form, but few attempts are made to really demystify it. Who came up with this? What is a standard deviation? Why are 68% of values within one standard deviation of the mean? The beginner is more or less asked to accept this as an article of faith.<\/li>\n
  2. Similarly critical values. You have to look up your answer on a table of critical values somewhere in the back of the book. Again the beginner is not troubled with any sort of explanation about where these critical values really come from. Why different scores for t, q, u, chi etc.<\/li>\n
  3. The internet is the enemy of the beginner. This is slightly unfair as there is some good stuff out there, but most statistics resources reinforce the above.<\/li>\n
  4. Statistics as a \u2018bolt-on\u2019 leads encourages surface learning. Statistics is a hurdle that has to be cleared. Technique is emphasised above understanding. Even the best teachers of quantitative social scientists don\u2019t have time or scope to get to grips with true understanding.<\/li>\n
  5. The pressure to learn statistical analysis software creates an additional barrier in time and learning.<\/li>\n<\/ol>\n

    Rebuttals and thoughts welcome.<\/p>\n\n\n