Quantitative data analysis: really kinky stuff or just standard deviation?

The following notes are made from Ch11, Analysing Quantitative Data, Tim Philips, in Social Research Methods: an Australian perspective. (Ed) Maggie Walter. Oxford UP, South Melbourne, 2006 (2008).

This is a very useful chapter if you need help in establishing a way of talking about the datasets. It uses examples that are almost exactly the same as our assignment.

Primary data collection: 3 very important things.
1 Time: Spend a sustained period of time designing the survey.
2 Resources: Attain a large research grant to pay for survey administration.
3 Specialist knowledge: Get advice from different experts about the best way to undertake particular steps.

Secondary data collection
When designing your own research instrument for high quality data collection is not feasible, undertaking secondary data analysis is an alternative. this generally involves the reanalysis or additional analysis of publicly available survey data., eg within a social science data archive, like ASSDA (see links at bottom of page). Supporting such an archive contributes to the shared scientific method standards of professionalism, transparency and integrity.

Looking at a dataset from the Australian National Survey of Social attitudes 2003,the textbook introduces a research problem for a mini-anlalysis: whether the difference in among Australians in feeling about globalisation is connected with divergences in patterns of mass media usage.

I put this in bold because I had to return to it, and re-read their background briefing. The textbook reports positive perspectives in the mass media about globalisation. It then produced evidence from a 2001 study which indicated variables such as education level, internet usage and overseas travel as factors towards Australian peoples’ cool feelings about globalisation. At this stage, the textbook example suggested bringing mass media consumption into the equation, which had been left out in the 2001 study. And finally, they chose a set of particular variables in order to create a null hypothesis for the cross tab analysis.

[the rest of this post is really boring - i'm just reiterating key concepts and recording some detail about the examples used, but it may not make much sense without the textbook]

1) frequency tables – good for initial feel of where responses are located across response categories. eg, around 20% or 1 in 5 responses relied on abc/sbs, radio and newspapers for their most news and information, but 37% relied on commercial tv channels.
Using frequency tables to calculate Summary statistics:
2) measures of central tendency -
“Individual variable data is measured at different levels, building in complexity from nominal data, through to ordinal data and interval data to ratio data.” (p292)
Mean: the average of the scores. Used when variables are measured at interval or ratio level
Median: the middle in a set of ranked scores. Used for ordinal level variables.
Mode: the most frequent score in a set of scores Used for nominal level variables
3) measures of dispersion -
Standard deviation – a statistic that shows the spread of scores or variance around the mean.
Variance – a measure of the spread of scores.
Illustrating central tendency: Median response for “Large international companies are doing more and more damage to local business in Australia” was Agree.
Median response for “International organisations are taking away too much power form the Aust. govt” was “Neither agree nor disagree”.
Illustrating dispersion: Among the questions included one about total number of years of education. Although the mean for men was 13.28, slightly higher than women (12.96), the standard deviation of men was more widely spread out at 4.10 than women at 3.86.
Cross-tabular analysis
So far, we’ve looked at a subset of questions gauging pessimism towards globalisation, and (b) a single question tapping preferred source of information. But we want to be able to bring these questions in isolations together so we can bear upon the question of key concern: is media usage connected with disengagement form globalisation, and if so what is the link.
1. return to the key question of concern, ie, what is going to be your independent variable that you’ll check against a range of demographic variables? For this example:
Independent variable: A7 Which of the following sources of information would you say you rely on MOST for your news and information?
Response choices: Commercial TV, ABC/SBS, Radio, Newspapers, Internet, Talkback radio, News magazines, Friends and family.
Cross tab analysis provides us with a frequency distribution within the categories of the independent variable.

End summary:
Quantitative data analysis is at its strongest when:
• looking at complex relationships (cross-tab analysis)
• making inferences from samples to populations (chi-square test)
• examining grand claims within social theory and specifying the conditions under which they hold up
• adjudicating between competing theories. NB: It should be made clear though, that while in some cases quantitative analysis may function as an ”objective”‘ perspective to view potentially competing theories, it may not be appropriate. For example, a cost-benefit analysis of Catholicism vs paganism may produce some interesting data, but is it useful?
Quantitative data analysis is at it’s weakest when:
• variables are poor measures of concepts
• the status of vairables in causal chains is determined arbitrarily
• statistical methods are used that place too many demands on the data
• it is done in the absence of theory.

Final notes:
• High quality quantitative data analysis always takes place through a process of iteration. This is where the researcher must be creative and dynamic.
• There is never simply a ‘right way’ to analyse a quantitative dataset  to derive an answer to a research question’ (Becker 1986)
• The research must figure out a persuasive and compelling analytic strategy. Try unleashing your ’sociological imagination’ (Mills 2000)
• A lot of it is about balancing convention and innovation. [sounds like design theory!]  Perhaps established ways of doing things are appropriate.  However, sometimes conventional approaches engender a sense of doubt within you (Bauman & May 2001), so you might be better off demonstrating your own version. In this way, imagination and intuition come to the fore as vital qualities for ’steering’ your data.  Giddens (1990)

Useful links
Australian Social Science Data Archive (ASSDA) http://assda.anu.edu.au/analysis.html
Nestar Light http://assda224-100.anu.edu.au/nesstarlight/index.jsp

References
Becker, H. (1986) Writing for Social Scientists. Chicago: University of Chicago Press.
Bauman, Z. & May, T. (2001) Thinking Sociologically. 2nd edition. Oxford: Blackwell.
Giddens, A.  (1990) The Consequences of Modernity. Stanford, CA: Stanford University Press.
Mills, C.W. (2000) The Sociological Imagination. 40th Anniversary Edition. Oxford: Oxford University Press.
Philips, T. (2008) ‘Ch11 Analysing Quantitative Data’. in Social Research Methods: an Australian perspective. (Ed) Maggie Walter. South Melbourne: Oxford UP.

a short lesson in qualitative research

Short, because I need to be writing notes on the quantitative textbook chapter.

This is a lesson i learned over the past weekend. In short, i got too close to my data. again. oh the pain. it was like pulling teeth with tentpegs. The lit review reared up and threatened to take over my whole life, a la that old honours crisis circa 2006. It’s horrible to see an old enemy again and just get flattened like that. My appalling time management also pitted itself against me and together, they were unstoppable.

So, how did I learn this wondrous fact? I felt the struggle. I felt responsible for the whole world, and everything that was wrong with the world could find its genesis in my research proposal. I am guilty of taking things to heart a little too much. Not that it’s a bad thing… I’m just recognising in myself the powerful attraction I have  to solving the world’s (ie mine) problems at every assessable opportunity.

And as terrifying as this was for me, I’d like to take a moment to thank each brave lecturer attempting to mark any work that had escaped the careful eye of dearest friends and lovers who’ve battled the odds and lent their proofreading and editing talents to my quest.

It doesn’t have to be this way! I can be a little less than perfect! (ok, a lot less)
If i think a little more abstractly,I can follow through many hypotheses,  rather than trying to push heavy concrete problems into some kind of tetris-shaped solution.

And in this attempt to care a little less, and think a little more, I am aiming to create some thing/s that are a bit more transparent, a bit easier to understand and engage with, for other people as well as myself. Instead of just feeling the weight of it all, and getting all mollified at the prospect.

that’s enough for now.

readings for week 4 57146

second boring post!

Hodge, G. 2000, ‘Knowledge Organization Systems: An Overview’, in Systems of
Knowledge Organization for Digital Libraries: Beyond Traditional Authority
Files, Digital Library Federation, viewed 14 July 2008.
<http://www.clir.org/pubs/reports/pub91/contents.html>.
Rosenfeld, L. & Morville, P. 2002, ‘Thesauri, Controlled Vocabularies, and Metadata’,
in, Information Architecture for the World Wide Web [online], 2nd edn,
O’Reilly, Cambridge, Mass., p. Chapter 9 Sections 9.1 – 9.2.3v

Reading for week 5 57146

Warning: boring post!

Readings for Week 5

Anderson, J.D. & Pérez-Carballo, J. 2001, ‘The nature of indexing: how humans and machines analyze messages and texts for retrieval. Part I: Research, and the nature of human indexing’, Information Processing and Management, vol. 37, no. 2, pp. 231-254 Particularly Sections 4 – 7

Mai, J.-E. 2005, ‘Analysis in indexing: document and domain centered approaches’, Information Processing and Management, vol. 41, no. 3, pp. 599-611 .

Note to self – write a short para on Elsevier experience.  Would I add it to ‘my databases’?

Ethics and research….hmm

Text book Exercise 3.1
What measures can researchers put in place to empower respondents?
➢    Give adequate time and information to make an informed decision about participation (work this into the Project Plan)
➢    Give participants the right to:
o    refuse participation
o    withdraw participation at any time
o    refuse any question
➢    Protection from acts which might cause harm or discomfort beyond that encountered in everyday life
➢    Allow some control over data by giving access to findings before publication and the right to veto certain types of presentation, ie who receives copies of the report and where it is published.
➢    Submit research proposal to an ethics committee or HREC.

Q2: What are the methodological advantages and disadvantages for bringing respondents into the research process?

➢    The future of your research (if what and where it is published) may become dependent on an individual or a number of people’s sense of shame or embarrassment.
➢    You could be handing over control to participant’s commercial / economic / emotional / sense of community needs.
➢    You will need to be able to communicate how your research will benefit the respondent, and or their community / workplace / environment or other context.
➢    Negotiating with respondents may help the researcher understand the respondents needs better and could potentially inform a more successful research design benefitting both the research field in question and creating more appropriate data to support social change.

Epiphany!

I know i shouldn’t really encourage such a subjective bar to set for blog posts, especially given my awol-ness of late, but this morning in the shower i had an idea.

I’ve signed up for a toolkit workshop at QueenSt Studios, a little artist run venture which supports various artists and performance makers by offering studio space, community connections and workshops for sustaining practice.  Like the one I’ve just started. Perma-culture – a hands on workshop looking ways and means of sustaining an arts practice. It’s going for 4 weeks, and it’s with Rosie Dennis. http://www.suture.com.au I don’t know why i didn’t tell more people about it. I feel a bit disconnected from old performance / queer scene communities, which is probably cos I’ve got my new plugs on, hooked into information management, library sciences and all that vocational turn stuff.

The epiphany! In Rosie’s workshop, we did an exercise which involved telling a story about a photo from your childhood. We did this from the perspective of 1st person, then from 3rd person (same photo), then incorporating some narrative techniques (thesaurus, repetition, delivery, alliteration) in each retelling. And then we retold the story with doubt, doubting ourselves at every turn, questioning each aspect with a view to seeing where this led us.

Remarkably similar to the question formulation of research design which I’m studying this semester. What is introduced as a tool for developing material for performance is also a key part of analysing research.  And this exercise in the workshop was fun! And I like having fun and being amused, things are less scary that way. So I’m really invigorated now about research design and how to focus a question with my syndicate group in class.

Design Journal #1 Getty images

Using March Chignell’s 4 criteria for usability to evaluate the getty images website. I’m looking for royalty-free images.

First impressions and a log:

Appears to be a fairly easy to use site, especially for an experienced/ web literate user. Simple bar menu at the top with pop-up menu made getting to the royalty free images quick and easy, at first. Hover on Image button, float down to”royalty free” images and click. New window. I’m not sure what I’m looking for, I’ll just browse – click on “see some of our top RF photos”. New window, slideshow or see all. slideshow is fast, click for detail doesn’t seem to bring anything up until I realise I have umpteen getty tabs open on the browser behind.

The pop-up “we’re here to help you chat with a specialist now” appears helpful at first, but it appears on each new tab of getty images that i’ve opened. I realise later that i don’t have to close this popup to close the tab. Still, seems like a lot of unnecessary clicking.

After browsing the “best of”, I think, oh, there’s a lot of straight men in these photos, I wanna look for the queers. Close window, return to main window, didn’t actually see the search bar at the top, scroll down to collections, pick a house collection at random.

Using various combinations of dyke, lesbian, queer, camp resulted in lots of photos near dams, some mono-gendered couples. “Camp” prompted a clarification window, which seemed useful, except the options (did you mean “camp” outdoors or “camp” summer camp? Neither!) were unsatisfactory.

However, first searches proved really unfruitful until i realised that the right sidebar had the rights-ready box ticked, instead of the royalty-free box which I was expecting. I think this is a major fault.

Chignell’s criteria:

Usefulness: I’m not sure if I’m asking if the design is useful or the website is useful? The front page is clean and attractive, the top menu bar with it’s drop downs contains at a glance most of the likely information one might be looking for. The coloured boxes in the main body of the page represent links to features that getty images wants to highlight.

Effectiveness: I think a website has to appear useful to be effective. Getty have recently changed their licenscing options and created simple presentations to explain this, in short grab boxes, an extended FAQ. Their information sends the user towards rights managed images if it’s for competitive advertising business or towards royalty free for creative types who just want to play around with images.

Learnability: Once I realised I had to check my search options before searching, I had better results. I discovered that royalty free had a more diverse range of images, which was what I was looking for, althought they weren’t as high quality as the rights-managed.

Likeability: I think I turned on getty images when I couldn’t find what I was looking for. Also, it’s a very smooth and slick design, it looks like a North American magazine, so I am a little distrustful of it’s usefulness for me, however advertising types selling brand packages for business in a tight deadline/ budget might find this an excellent site.

reorientation//reintroduction

This time I’m actually gonna use this blog! I have two new subjects this semester, and my major overarching goal this year is to keep my study under control, not let it blow out and take over my life. To do this, I’m gonna use endnote to organise my reading lists as well as references, I’ll have a weekly planner with regular slots for reading, revising and working on assignments. Each subject also has a journal type assessment, which I’ll keep on this blog.

First subject: information architecture and design.Woo! Lots of user-centred design process stuff, for print and online ‘publications’. I expect to get quite critical about relations between things and the people who use them. Keeping a design journal for this one.

Second subject: discovering and accessing information. Otherwise known as information retrieval, I expect this class to be an intense ramping up of online search skills and getting my head around librarian systems of organising information. Keeping a portfolio of learning activities for this one.

Ok. down to work!

flatworld

Our books are free online. We offer convenient, low-cost choices for students – print, audio, by-the-chapter, and more. Our books are open for instructors to mix, mash, and make their own. Our books are the hub of a social learning network where students learn from the book and each other.

reblogged from Harold Jarche

Starting afresh

And wondering how i can ever explain this recent absence, and sudden about-turn.

Nevermind. I’m getting all vocational and turning towards the libraries for some kind of career-like thing. And currently boggling at the reading lists. It’s funny how these subjects in information management have such long and intense readings, yet seem to require such a different way of reading them. No more the read and savour, thrill and inspire of old performance and gender studies! It seems to be all about skim reading looong e-readings. And it will take a few more weeks of learning what not to print too. The combined effect of the subjects I’m studying seem to place books (as opposed to journals and online content) on the very bottom of the reliable sources of information.

Hopefully it will encourage me back into the writing.