Thursday, June 28, 2012

Working with multiple data

Attended a free webinar this morning via QSR the company that sells nVivo. The topic was on public consultation research presented by Patrick O'Neil rom Lincoln University and Dr. Lyn Lavery from academic consulting. Introduced by Kate from QSR.

This is the last of a series of 3 on working with different data came about through the public consultation process Chch city council ran after Chch earthquakes. with previous webinars available on research preparation (for the unknown) and rapidly synthesising and analysing data.

The projects was focused on the Share an idea project used to inform the Chch city plan. 14thMay and series of interactions also used with all data collected by July. 2 teams involved with city council And overseas consultants. Coding structure was set up to deal with multiplicity and amount of data. Needed to ensure data collected could be collated to be meaningful. Topics set up and sub themes emerged. Important to have shared understanding of what sub themes mean to all researchers.

All data, regardless of type, coded to nodes and subthemes on nVivo. large amount of data had to be analyzed within a short time frame. Need to ensure all data was analysable. Data gathered through expo, website, council forms, workshops.

Website collected tweets (over 20,000), emails, group ideas. Council forms included letters, emails, voicemails. Expo included post it notes, pictures and drawings, focus group transcripts etc.

NVivo allowed word, excel files, PDFs audio, video and images with all could be coded to the same topics and sub themes. Plus 5 researchers could all work on data at the same time using nVivo server.

Examples of data types from expo then described.
1) Let's hear it used written information on a form - with 4 questions - what missed, what to retain, how to make Chch city better, what is most important About vision for Chch. Transcripts entered as answers to questions and then coded to sub themes.

2) Web tweets with 4 themes - move, life, space and market. Engaged people before expo and then others could see their own and others and provide feedback and stimulated more ideas. Above coded with tweeters age And suburb and tweets coded into sub themes.

3) Post it notes from expo generated 17,000. Entered into an excel file and manually coded to theme. Then text search used on excel files deployed to create nodes. Merged into existing nVivo data. 1/3 of post it notes then coded using coding density functions.

4) YouTube videos used to capture interviews up to 3 minutes with people who preferred to provide oral feedback. Video transcribed within nVivo and again coded to sub themes. Selected parts were transcribed by firstly doing a first watch and then only transcribing relevant parts.

Children's art to help children articulate their ideas. Photo of art uploaded to nVivo, regions of the 5) picture can then be coded by identifying on the picture and labeling.

Outputs also used nVivo visualization tools and can be run using queries to generate tag clouds ( word frequency with stop word list to remove unnecessary words) and word trees (to illustrate words that are searched and other things associated to each word).  Visualisations could be used to quickly display the most common ideas to inform good mass communication and quickly represent a particular topic and display a large amount of content.

All in a good opportunity to have an example of how to better use nVivo for qualitative analysis. Plus information on nVivo 10 which allows for social media data to be more easily used plus need for nVivo to be able to handle very large projects more efficiently.