Instant Ideas and Collaboration
Our next #uklibchat on Tuesday 5th May 6.30-8.30pm (UK time) is titled:
Stats: What are they good for?
(Agenda will be out soon)
In keeping with our topic, our featured article is by Mathieu Lubrun about his use of statistical software to analyse qualitative data for his workplace, GSM (Greenwich School of Management) London.
In order to offer a better service to the students at GSM London we asked our students to answer questions about their experience as students. The questionnaire is a yearly occurrence and asks many specific questions as well as a part where student can leave non-specific comments about whichever subject. Asking non-specific questions about the quality of the service provided can be very helpful in finding out the issues that might not be covered in a questionnaire. After gathering around 2500 answers I proceeded to find an appropriate software to make sense of the data.
Despite the accessibility of a tool like Survey Monkey to create surveys and gather data about user experience, analysing the information obtained through this means is still complicated when it comes to obtaining information from qualitative, rather than quantitative, data. Survey Monkey only creates a cloud of the most used words, which can be useful if your questions are focused on specific aspects of the library services, but not when you have to analyse a survey that dealt not only with the library, but with the school’s performance as a whole.
I tried to use SPSS, a software developed by Microsoft for statistical analysis whose licence was available at the school, but soon realised it would be useless for qualitative data. Indeed, word based answers can only be entered in SPSS if they are turned into numbers. For example, if your survey contains a scale with grades such as Agree, Somewhat agree or Disagree, these answers can be transformed into 1, 2 and 5 respectively. However, a complete sentence will be treated as one single item and not as a collection of items related to each other.
After I came to this conclusion I turned my attention to software such as Nvivo and Leximancer which are used for sociological analysis of qualitative data. Although Leximancer seemed very efficient, it appeared too high level for the kind of basic analysis I wanted to do. Nvivo on the other hand offers a perfect solution because of the simplicity of its design, which resembles Outlook, but also of its adaptability to many types of data (Excel files, videos or data gathered from social media). Even if the data I had available was only on an Excel file, it is important to think of what future questionnaires and method of data collection could be developed with such a tool.
In this case, the selection of the software to analyse the open comments came after the design of the questionnaire, but for future surveys it will be essential to think about the methods available to analyse the data. This factor will then be essential to determine what sort of question to ask and what kind of data to gather (video or audio recordings of interviews, comments from social media, open ended questions).
Through Nvivo I was able to find the most important issues student were facing in the library by doing research on the word ‘library’ to see which words were the most commonly associated with it. For example, the numerous demands for a larger collection of books was one of the comments most frequently left by students and will inform our acquisition policy. This information would not have been picked up by any of the other software and would have required a fastidious reading of every comment. Thanks to a more appropriate software, a simple word query gave me the number of mentions of the word book and a list of every comment where it was mentioned and in which context.
Although a tool like Nvivo can make data analysis much easier, it is important to understand that you still need to engage with your data to make sense of it. Basic queries on Nvivo will push forward specific issues but it will not make sense of it for you.
In my opinion, surveys and questionnaires should be at the root of decision making but can only be relevant if they are done on a large scale. The analysis of the data gathered, especially open comments, can then help to discriminate between issues that might seem relevant to librarians but are irrelevant for student. Furthermore, a school wide survey offers the opportunity to get feedback from the student who do not often go to the library and to understand what needs to be improve for the library service to be more inclusive.
Lord of books