Affinity diagram method VS expert reading of data

7 Dec 2010 - 3:51pm
3 years ago
5 replies
1191 reads
Ali Naqvi
2008

Hello,

in order to analyze contextual inquiry data either affinity diagrams can be created (together with a group of people) or an expert reading of the data can be made. 

I am the only user experience consultant hired so far and most of the other consultants are hard core developers. I have decided to analyze the collected data, in a more traditional kind of analysis. (expert reading)

When more UCD practitioners are hired, I would like to start using the affinity diagram method. 

I would like to know whether you guys use this method, and if there is something I or the team in general should be aware of? Any drawback? How can we make full use of it?

Regards

 

Ali

 

Comments

7 Dec 2010 - 6:05pm
Michael Simborg
2009

Inconext - rapid contextual design and contextual design

wandereye

On Dec 7, 2010, at 15:22, Ali Naqvi wrote:

> Hello, > > in order to analyze contextual inquiry data either affinity diagrams can be created (together with a group of people) or an expert reading of the data can be made. > > I am the only user experience consultant hired so far and most of the other consultants are hard core developers. I have decided to analyze the collected data, in a more traditional kind of analysis. (expert reading) > > When more UCD practitioners are hired, I would like to start using the affinity diagram method. > > I would like to know whether you guys use this method, and if there is something I or the team in general should be aware of? Any drawback? How can we make full use of it? > > Regards > >
> > Ali > >
> >

7 Dec 2010 - 9:05pm
Chauncey Wilson
2007

Hello Ali,   Affinity diagramming is used by many, but there are some drawbacks and best practices.  Many of these would also apply to an individual analysis of qualitative data.   1.  Data from contextual inquiry (or other qualitative methods) needs to be separated into separate "semantic units".  To break up the data, you need to develop a consistent set of rules.  For example, if you have notes that express mutliple thoughts, you might need to indicate context in brackets so the meaning is understood out of the flow of the notes. Some of the books on qualitative analysis describe how to break text data into units that can be words, phrases, sentences, or groups of sentences.  You can develop a set of rules as you work with the data and then follow them consistently.  Each note should express one idea or statement. 

2.  Number (or code) all your notes so you you can put them in a database and track them months after you collect the data.  Contextual inquiry data can be used to generate use cases, tasks, scenarios, errors, job roles, objects, attributes, etc.  You can number your items and also tag them with some of the general terms.  For example, your users might discuss artifacts and objects that they encounter in their work and tagging those would allow you to quickly look at all the objects (or attributes) that emerged from your data.

3.  When you convert your text data into notes, keep the language of the note as close to possible to the original language.  For example, if a person uses the acronym, ASA, include the acronym and then explain it in brackets so you can differentiate your input from your participant's input.

4.  You might want to track your notes by user. 5.  Keep in mind that cognitive biases can influence how qualitative data are interpreted.  For example people tend to weight more memorable events as the most important (whether they are or not).  This is known as the availability bias.  There is also the tendency to overestimate the representativeness of small samples of data.  You can review some of the cognitive biases by looking up "cognitive bias" in Wikipedia.

6.  One output of an affinity diagram is a list of insights.  Usability experts Mary Beth Raven and Alicia Flanders suggested creating an insight sheet for an affinity diagram that describes the insight, an example to support the insight, and a design implication or idea to consider.

7.  David Siegel and Susan Dray have some great tips on some sample slides from a CHI 2003 session at http://www.chi2003.org/docs/t36.pdf 8.  There is an interesting tool that you might find helpful called Concordance.  It does word counts and then provides context around particular words.  It is inexpensive and is a way to get insights if you can pull all your data into a simple text file.  See http://www.concordancesoftware.co.uk/

9.  Consider putting your data into a spreadsheet (or database if you are proficient with something like Microsoft Access) and tag your data in several ways.  For an individual review of qualitative data, tagging might be a powerful way to examine your data.  You might have tags like:  "breakdown", "goals", "Object", "attributes", "collaboration", "error", "performance", "design idea", ......  

10.  The degree of domain knowledge you have will very likely influence how you organize your data.  You might invite project managers (PMs) or someone with domain knowledge and knowledge of users to help you. 

  Thanks, Chauncey        4.    

  On Tue, Dec 7, 2010 at 4:17 PM, Ali Naqvi <Ali@amroha.dk> wrote:

Hello,

in order to analyze contextual inquiry data either affinity diagrams can be created (together with a group of people) or an expert reading of the data can be made. 

I am the only user experience consultant hired so far and most of the other consultants are hard core developers. I have decided to analyze the collected data, in a more traditional kind of analysis. (expert reading)

When more UCD practitioners are hired, I would like to start using the affinity diagram method. 

I would like to know whether you guys use this method, and if there is something I or the team in general should be aware of? Any drawback? How can we make full use of it?

Regards

 

Ali

 

(((Please leave all content below this line)))
____
8 Dec 2010 - 3:31am
Ali Naqvi
2008

Hello Chauncey,

thank you so much for replying to my post! Your post was of great benefit and the links were useful.

In order to analyze qualitative data, I use Steinar Kvale's methods. (A Norwegian professor with expertise in Qualitative Method Development)

I used his methods in my thesis http://amroha.dk/thesis.pdf , please look at page 6 under the Methods section.

I believe that his methods are good in filtering out cognitive biases.

Regards

Ali 

8 Dec 2010 - 12:06am
Adrian Howard
2005

On 7 Dec 2010, at 21:17, Ali Naqvi wrote:

> Hello, > > in order to analyze contextual inquiry data either affinity diagrams can be created (together with a group of people) or an expert reading of the data can be made. > > I am the only user experience consultant hired so far and most of the other consultants are hard core developers. I have decided to analyze the collected data, in a more traditional kind of analysis. (expert reading) > > When more UCD practitioners are hired, I would like to start using the affinity diagram method. [snip]

I wouldn't immediately dismiss doing affinity diagrams with the developers - or any other folk involved with the project. A second point of view is always handy - even when it's from a non-"expert".

I regularly do this sort of thing with developers when I'm the sole UX person involved in a project and find it useful.

I also find that doing these sorts of exercises with developers greatly helps with communication about the project's design as you move further into the product development process.

Cheers,

Adrian

8 Dec 2010 - 3:35am
Ali Naqvi
2008

Hello Adrian,

Isn't it true that in order to do an affinity diagram with the developers, these developers need to be present during the interviews or field observations? How else will they be able to contribute? I am not sure whether the hardcore developers will be attending the interviews/field observations.....

I need to talk to them about this....

 

Ali

Syndicate content Get the feed