PhD

DevonThink Pro for PhD

by

Well I have been promising for a while to write a blog post on DevonThink.  On the details on how to use it and manage information.

This… is not that blog post.  This is an interim short post on DevonThink and why I replaced Evernote with DevonThink on the Mac.

But just a quick note to say that I do use DevonThink a lot in managing my phd research.  

Some highlights for me include:

  • Tagging of documents (I tag by author, journal, year, keywords, and reference [James & Huntley 2009]).
  • Seamless working with Preview so I can immediately open up a PDF and annotate it (highlights) straight into the database (saves back to where Devonthink does
  • Full text indexing of the OCRd PDF (and the fact that you can then copy the OCR’d text out of the PDF into a separate ‘quotations’ folder.  This is a DevonThink Pro feature.
  • I use smart folders to search quickly for the tags etc that I want (created several smart folders that you just double-click and edit the search criteria e.g. tags).
  • I have two folders – one for the original PDF documents, and another for quotations (I copy the quotation out of the PDF with page and paragraph reference, and put that in a comment when I cite it in my phd thesis).
  • Oh – for quotations, I tag it with the original paper, tag the quotation with any relevant keywords, and put the page order/para order in the title of the quotation (e.g. “p299 para3 Jones & Huntley (2006) provide a theoretical basis for the anomalous findings of Blake & Waldron (1976).”  That way when you do your search if you have multiple quotations from the same source you can see them in order and, perhaps a little, context.
  • The magic-hat searching is quite useful once you build a bit of a database up of papers – click the magic hat and you’ll see papers and quotations the AI thinks are relevant inside your database.

I am also exploring the RSS feeds from journals more, but they do seem very useful as the magic-hat will tell you if there are new publications (within reason) related to your topic area.

So… not the full post on DevonThink that I intended to create but this is a good potted summary of what matters to me.

Only negatives:  data is only in one place (doesn’t sync to multiple devices like EverNote – and no DropBox doesn’t count).  And Mac only – I am concerned that if I move platforms I won’t have access to my DevonThink work any more (files can be easily moved but the tagging I’m not so sure about).

Academic Writing Tips that work for commencing RHD and PhD students

by

Yesterday I reviewed a doctoral consortium paper for a fellow phd student here at UQ.  I thought I might as well put some of the comments I made up as a blog post.  I want to be sure I just document up some tips or advice for students writing their first doctoral consortium paper.  For some people the below will be old hat – for others, not so much.  As an early phd student you need to develop these skills.

Firstly be clear as to what your paper does and what the introduction needs to do:

I believe an introduction does the following:

– Attention grabbing statement of the problem.

– Mini-literature review

– Identify the next steps in the research area (motivation), and identify the scope of the research question

– Quick overview of research method

– What are your contributions?

– Roadmap to the rest of the document. I feel your introduction is too long (1page out of 7 and 591 words, which then cuts into the rest of your research proposal.

Caveat – your discipline may differ.

Don’t over-cite:

If a long list of papers that are only cited once are in the introduction without explanation, I think you need to cut them down for the introduction.  Unnecessarily blows out your reference list.

Which vs that:

Which should only follow a comma (kind of a golden rule).  P88 Strunk & White – go which-hunting ☺.  That refers to a specific item, which is for a class or non-restrictive group.

Have transition paragraphs at the beginning of sections – a style thing:

I think the role of this transition paragraph is to identify what is coming (“we will talk about this, then that, then this, and finally we can move to the next section to discuss X”). I feel this may be a bit long for this role – I think you should flag what is coming, but not describe how it works.

Another writing tip is to avoid long sentences.  And one-sentence paragraphs.

Don’t over-rely on the words of others:

Is the long quotation from this paper really required?  Takes space away from your own work.

Don’t use ‘it’ as it creates needless ambiguity.  Like one-sentence paragraphs, it’s a no-no.

Be careful with ‘because’:

Never start a sentence with ‘because’ – did Mrs Sandilands not teach you in Grade 2 ☺.  Maybe try ‘As’.

If you have a research model you have developed with labels, be sure that your discussion is consistent with the labels (i.e. section headings).  Also try to follow the same order throughout the paper.

Also be careful with word use:

Unless you are very clear on its meaning, a phrase like ‘unit of analysis’ is inviting trouble.  Use ‘focus’ e.g. “user experience is the unit of analysis” is better off as ‘user experience is the focus’.

Similarly confirm vs validate vs explore

An exploratory study explores your research model.  A confirmatory study confirms it.  Don’t’ have an an exploratory study validate your model, and don’t have a confirmatory study validate.  And don’t have a confirmatory study further validate the validatory study.  Does it ‘validate’ or ‘confirm’ ☺. I’d lose ‘further’ as it sounds like you validated it once, and you’re now validating it again, and it makes no sense to validate what is already valid. However, you can confirm a valid model.

It’s a bad idea to finish a writing section on a dot point:

You need a concluding paragraph here so you don’t finish on a bullet point.

In the conclusion:

I think you need to state earlier and more clearly, what are the practical and theoretical contributions? Theory first, practice second. I don’t see any contributions to practice identified?  Have a strong final concluding sentence – hopefully that reflects the theme of your strong opening sentences.

See I can blog occasionally.  The above tips are not everything, they are just the comments that I made and this is just a grab-bag of points.  As some readers may know I have written a template in Scrivener for academic writing (see http://michealaxelsen.com/blog/?p=839) that formalises some of these comments into a structure.  Although I’ve had struggles with Scrivener (http://michealaxelsen.com/blog/?p=2930) and I have to say I still haven’t come completely to grips with tables, I have found Scrivener to be a really useful writing tool.  I’ll probably change my mind again by morning tea as I battle tables.  

Post-Script:

Having just submitted a doctoral consortium research-in-progress paper, my supervisor was at pain to give me some solicited feedback on my own work :).

Don’t use words like innovative research design, significant contribution, or novel approach – such value-laden judgments are best left to the reader.  Leave these adjectives at home :).  

Thanks:  Micheal Axelsen

AMCIS 2012: “Continued use of intelligent decision aids and auditor knowledge: qualitative evidence”

by

So, I submitted a paper to AMCIS 2012, an academic conference to be held in Seattle this year (http://amcis2012.aisnet.org/).  The paper was accepted (“I also think that this should generate some interesting discussion and hopefully receive further guidance to help the authors publish their work in a journal”) and so I am off to Seattle in August.

This paper was written solely by myself, without revision by supervisors, so I am quite happy about that.  It is also based on my phd, which is very helpful, and of course reviewer comments are very good to help with this process.

Anyway as I haven’t blogged in a while – here is the abstract of my paper “Continued use of intelligent decision aids and auditor knowledge:  qualitative evidence”:

The Theory of Technology Dominance proposes that continued use of intelligent decision aids (IDAs) relates to a decrease in auditors’ decision making skills, or deskilling. Prior research has considered deskilling in terms of auditor declarative knowledge. This research considers deskilling in relation to auditor declarative and procedural knowledge through an extended research model. A novel, rigorous and repeatable qualitative research method using automated text analysis (Leximancer) is developed for the analysis of significant bodies of text. Nineteen senior auditors in three audit offices were interviewed, and the transcripts analyzed. The findings indicate strong support for the hypothesized negative relationships between three constructs (the extent an IDA performs routine and time-intensive tasks, the dependence of an auditor on the IDA, and the auditor’s time with an IDA), and an auditor’s declarative and procedural knowledge. The results indicate avenues for future research, and provide guidance to practitioners in the use of IDAs. 

Once it is published I will put a link to the paper on this blog entry. 

Emerging Online Social Communities and Emergency Services in a Connected World

by

The following was a ‘joke’ abstract I wrote on Facebook (yes I know – get a life), and was inspired by a Queensland Police Service Media posting in relation to a specific incident.

When you are writing a paper I always finds it helps to write the abstract first – at least to clarify the purpose of the paper.  Obviously, you rewrite the abstract later.

This study develops a theoretical model from a comprehensive literature review to identify factors relating to the successful building of online communities. The developed theoretical model is then considered in the important context of emergency services and their extension to the online world. The theoretical model undertakes a secondary data analysis of material published in the public domain by these public service entities in light of the theoretical model. This qualitative analysis is supplemented through consideration of quantitative measures of the success of these communities through espoused key performance indicators such as number of ‘followers’, community engagement (measured as percentage of respondents sharing, liking, and commenting upon announcements), and time to resolution.

The mechanisms of success are explored through the qualitative analysis of the material and depth of comment made through publicly available social media. This study contributes to the practical understanding of public sector emergency services response in an online environment, and furthermore provides a theoretical framework for the consideration and extension of social media effectiveness in the current integrated social world.

Unfortunately QPS delete some of their posts when there is no longer a public benefit arising (e.g. they found someone they specifically named) – which makes getting some of the numbers a little more difficult.

I was reminded of this because of this call for research students at QUT:  http://networkedblogs.com/wvQqn

Mac OSX, Scrivener and Word for Academic writing

by

Well, about 18 months ago I bought a new Macbook Pro.  I had sworn that Mr Gates’ hegemony had to deliver on Windows 7, and, although it was OK, it was not quite to the standard I wanted.  So I thought that, as I am currently doing my phd full time, it would be a good idea to try out this Mac thing full time. On account of how my need for my computers to play well with others on a corporate network is less.

And, speaking as a Windows user of long standing,  generally it’s been a Good Thing.  Sure, the Macbook was expensive compared to the Windows equivalents, but then the software is generally cheaper and OSX just feels more bulletproof.  Of course, software development is a bit harder to get into (but then AppleScript is very good), but at least when Office 2011 came out they brought back VBA for Mac Office.  Thank goodness!

And there are two items of software on the Mac that I have coveted and used a lot since converting to the Mac side. Devonthink Pro is one that I have just not managed to find an equivalent for – which should be the subject of a future post – and Scrivener is the other.  In fact, I made a video on YouTube and a post here on my blog, and since it’s consistently my most popular blog post and video by a mile, there’s a lot of love for it.

So, I have adopted Scrivener for writing.  And it’s a great all-round writing package.  The structure, the mechanism behind it, and the ability to compile to many formats (including ebook) and slice-and-dice your writing is wonderful.  I of course also have Office 2011 for ‘other writing’ and for final formatting.

The default for most phd candidates is of course Word, and on the Mac that’s periodically been a beast competent at short writing but not so good at long writing.  However, I have recently been using it to write a large report for a client (55 pages, 20,000 words)  and in the space of 24 hours it lost me 6 hours’ (yes, even with auto recovery files being saved every 10 minutes).  I would be editing and moving text around in Microsoft Word and then I’d get the spinning wheel of death (see – I have turned into a Mac user). Word would just end up not responding.  And when you’re consulting for a client that is paying hourly, losing that many hours is not much fun for your hip pocket. Nothing like the market’s invisible hand to make you think about your software choice.

Turns out there is a bug with the ‘Smart Cut-and-Paste’ function – go to /Word /Preferences /Editing Options and turn off Smart Cut-and-Paste and that problem with Word in Office 2011 crashing should go away.  After I found that tip I proceed to write for a week without a single crash.  I had been blaming it for EndNote incompatibility but this document wasn’t using that and I’d disabled that functionality.

So.  Word on the Mac can be made to work.  Although you can still lose work, and it is dependent upon you regularly saving your document if you want to avoid wholesale problems (and saving is slow on Word).

And so to Scrivener.  I love Scrivener.  Mostly.  The structured approach is good and the ability to move your text around is fantastic.

However – there are several issues for academic writing.  Graphics were a problem for a while but I worked out you need to draw your diagram in another package (e.g. Powerpoint), save to PNG, import the image, and then scale the image to consistently be the width of the page (411 pixels, incidentally).

Firstly – and insurmountably – you do have to play with the world of others if you are a co-author with people.  If you are the primary author, then you can compile, send it to your co-authors, and have them mark it up with changes for you replicate in your document.  But it’s a bear if your supervisor says, ‘Just give me the electronic copy and I’ll finish the report for you’.  Particularly since styles and text tags (such as the <$n:table:demographics> table you have on page 6) don’t translate to tags in Word – when they are expecting these things.

Secondly – and I suspect uniquely to academic writing – Scrivener’s tables, well, there’s no better phrase to describe it:  Scrivener’s tables suck.  Mostly because if it’s anything other than a straight vanilla table with no merged cells and so on, aligning tables is pedantic.  And I know that that’s because it’s really OSX’s tables.  And I know that Microsoft Word’s tables are proprietary and thus evil.  But, the tables just cannot be made to work.  I had gotten to the stage where I would do the tables up specifically in Word, and then import the RTF document (putting all the tables in their own sub-directory, with the same name as the $n:table reference in Scrivener) into Scrivener.  And that sort of worked when it was compiled out.  But sort of not.

Thirdly, a lot of academic papers come with arbitrary word limits and requiring submission in Word.  I have just had a paper accepted at AMCIS that was ‘5000 words including tables and references’.  Great – so I have to compile with EndNote every time I want to know how many words this thing is (noting, I suppose, this is really a function of EndNote’s sulkiness and thus you have to do this with Word too).  But you end up compiling from Scrivener to RTF, formatting EndNote tokens, saving to DocX and then reviewing the final output and changing a word on page 4 – and now you have two copies of your paper.  Sigh.  So you track change that paper, and then re-input your changes back to Scrivener later… which is every bit as painful as it sounds.  And if you then edit the paper in Scrivener, you’ll have to re-do all that formatting again.

Finally – it is in the nature of the academic to procrastinate.  If a paper is due at 5pm, then at 4:55pm you will be madly checking the format.  One of the great things Scrivener does is separate the writing from the formatting.  But the formatting is integral – presentation is communication.  Formatting the paper can take Quite a While.  And as Styles in Scrivener don’t work as they do in Word, a decision to make your section headings 18 points instead of 16 points can really wreck your day.  This is a problem when you are up against a tight deadline.

So… I am conflicted with Scrivener.  I think that for writing the Great Australian Novel it is the perfect tool.  And being able to compile different versions of the same document is excellent as well.  However, there are a few issues that I am trying to get my head around to make it work for academic writing. I just may have to accept that I do not have the time to do that and thus may need to ‘stick with Word’ for my PhD.  As a bit of a ‘new tool’ buff, and being able to see the advantages of Scrivener, that hurts, but more procrastination just isn’t an option when you are past the journey to knowledge and wanting to graduate.

[Footnote:  Since writing this post, I have gone back to using Scrivener as Word crashes far too often on the Mac.  Word is fine for two page letters on the Mac but that’s as far as I’d trust it.  I have written an automatic style formatter in Word (very rough, very basic, and very ugly) that converts fonts of a certain size to a Heading 1, 2, 3 and 4 style in Word – see this post here:  Scrivener to Word Visual Basic Fromatter].