Software Choices for Survey and Interview Data

An unexpected part of this project was the need to familiarise myself with a range of software in dealing with various parts of data collection, storage, and processing.  It turns out that each method in a mixed methods project presents its own set of requirements.  Below is an overview of the software I decided to use and how it all went.

Data collection took place in two waves.  The first was an online survey and the second was a serious of semi-structured interviews.  For the survey, I decided to use Unipark.  In comparison to other options, I found it to be affordable, customisable, and well up for filtering.  It is perhaps, a little over-engineered, and I did have to rely on the how-to videos and user manual more than I would have liked.  But happily, both were available and comprehensive.

In order to contact and keep in touch with participants, I found VerticalResponse provided a cheap, reliable mail managment service that kept my mails from getting caught up in spam filter.  It allows contacts to unsubscribe if they wish (but it doesn’t let you ask them why unfortunately), and it also (crucially) lets you know if your emails bounce.

For the interviews, my main tool was my iPhone which has a great voice recording app.  However, participants also created personal network maps with VennMaker.  Again, it’s very customisable, which is brilliant.  You can design a quantitative personal network interview if you like, or use it in the interview as a drawing tool.  It did the job and sparked some very interesting discussions about relationships, but some participants found it tricky to use.

Because I work from home and at work, using both a PC and a mac, secure storage has been a tricky issue.  Zipping and unzipping files made for a lot of copies of things that I needed to be sure to keep up with deleting, and a lot of encryption programmes work on a PC or a mac, but not both.  In the end, I settled on storing data on Dropbox and securing it with TrueCrypt, both of which are free.  TrueCrypt secures the file within Dropbox, so you need your password to open it whatever computer you’re on, and no, even Dropbox can’t open it.  Helpful instructions on getting set up are available here and here.  This has been a mostly painless system.  The only hiccup I’ve had is when I’ve worked for too long without saving the active file to the TrueCrypt file.  If you leave it dormant too long, TrueCrypt closes the file of its own volition and any unsaved data will be lost if you close the programme you’re working on.  To make sure this doesn’t happen, just check that the TruCrypt file is still active (or mounted in TrueCryptSpeak) before you close the file up.  Update (29 May): TrueCrypt has just shut down and declared itself unsecure, so it’s back to the drawing board on security.  Update (26 August): The lovely open source community of True Crypt Next  is auditing the True Crypt software.  They recommend continuing to use True Crypt pending results of the audit and have downloads available on their site.

For the statistical survey data, I’ve relied on SPSS for analysis.  It’s a handy gateway for a statistical noob, although it is very expensive if your institution isn’t shelling out for licenses.  Acton and Miller’s textbook is very helpful for getting started.

For the network data, I’ve tried Pajek, Gephi and Ucinet.  I started out on Pajek, mainly because it’s free, and there is a great textbook that walks you through the programme and social network analysis at the same time.  Gephi is also freely available, and it creates much more presentation-friendly images than Pajek.  However, it’s not quite as powerful in terms of analysis.  I recently decided to shell out for Ucinet, and I haven’t looked back.  It takes a bit of getting used to, but it’s a powerful programme that runs more kinds of analysis than I know what to do with.  Again, there’s a brilliant textbook available that links up the concepts and the programme.

For the interview data, F5 (or F4 for a PC) has made transcription much easier.  It adds timestamps as you go, linking the transcript to the recording, and it allows you to adjust the speed of the recording.  Shortcuts also speed up the process, letting you type only a letter or two for common words and phrases.  At 25 euro for the educational package, it’s a bit of a spend, but worth it, I think.  The plan is to use NVivo for coding, but I haven’t quite gotten there yet.  NVivo also has some transcription capabilities, so if you’ve got a license for it, there’s probably no need to buy extra transcription software.