Version 8.1 for Windows and Mac comes with a new analysis tool – global filters. Global filters are a powerful tool to analyze your data. As compared to local filters that you can invoke in each manager, global filter have an effect on the entire project. If you set a document group as global filter, also the results of the code-document table or the code-cooccurence table will also be effected. For example, if you cross-tabulate a code category ACTION with another code category OUTCOME, you can see which actions are related to which outcomes for all of your respondents. If you set the document group: ‘gender: female’ as global filter, the results of the table will change only showing which actions are related to which outcome for all female respondents.
In a code document table, it allows you to combine two variables without having to create a smart group. Given the above example, a code-document table including all ACTION codes by document groups “have children”, “do not have children”, the results will show the quotation frequency of all action codes for ‘female respondents with children’ vs. ‘female respondents without children’.
In network, all entities that do not pass the filter criteria will be faded out. Or, if you select the ‘import neighbor’ or ‘import co-occurring’ options, you can control the process by only importing entities that pass a given filter criteria.
An additional benefit is that you can focus your analysis on certain aspects of your project. All selection lists become shorter and if you know that you only want to work with 5 out of your 15 categories for the moment, you create a code group that only contains the codes that you want to work with at the moment and set this code group as global filter. Similarly you can create document groups if you want to focus on only a sub set of your documents for a given analytic task.
Currently only groups can be set a global filter. Other entities will be added in the future.
Creating a global filter means creating a group:
If you want to set a single document or code as global filter, you need to create a group for it. This is a current work around until it is possible to also set single entities as global filter.
Global filters can be set in each manager. Right click on a group and select the option Set Global Filter.
Similar to the visualization of local filters, global filters are also indicated by a colored bar.
The choice of color at the moment may look arbitrary from a Mac user perspective – but soon the Mac interface will also be brightened up and will include colored icon as in the Windows version.
As mentioned above, document group filters have an effect on quotations. If you open the Quotation Manager and a document group is set as global filter, you will see an orange global filter bar on top of the quotation list.
In Figure 4 the global filter is combined with a local code filter (beige colored bar). This results in the following query: Show me all quotations of the code “reasons for nhc: self-centered” for all female respondents.
Figure 5 shows an example where a global code filter has been set for a group of codes that is relevant for a specific analytic task.
Figure 6 shows a network that has been created with the help of a global code filter. The question behind the network was: Which positive and negative effects of parenting have been mentioned by those with 1, 2 or 3 children.
These were the steps to create the network:
The data are comprised of 140 comments to a parenting blog and a NYTM article Thus, there are many comments by different people in one data file. Therefore demographics like gender or number of children was coded.
In order to filter out all other aspects, a code group was created that only contained the three attribute codes (#fam, 1 child, #fam: 2 children, #fam: 3 or more children) + the effects of parenting codes. Subsequently this group was set as global filters:
The next step was to open a network for each attribute code and to import all co- occurring codes. As the global filter was set, only the ‘effects of parenting’ codes were imported.
These were linked with the attribute codes:
Based on these, a network with all three attribute codes was created. This was achieved by adding all attribute codes for number of children into one network and importing all code neighbors:
Figure 10 shows the resulting network using the organic layout and poly-line rerouting option:
Next the global filter was changed to codes from the code group: ‘Positive effects of parenting’. All other effects are faded out in the background.
Figure 12 and 13 shows two code co-occurrence tables. The first one shows the various opinions about the relationship between children and happiness of those respondents with and without children that commented on the parenting blog. The second table shows the same for those commenting on the New York Times Magazine article. If you deactivate the filter, you see the results for the entire data set. This applies to the table as a whole and to the quotation list that you get per code at the bottom of the window.
Figure 13 and 14 shows two code-document tables. Here the reasons mentioned for having and for not having children across different educational levels and gender are compared. In addition a code group filter has been set that contains codes that are only relevant for the research question considered at the moment. The comparison by gender in addition to educational level was facilitated by setting a document group as global filter.
Given the above example, if you want to see all quotations coded with “reasons for nhc: self-centered”, you click on the check-box for the global filter.
If you change a global filter, the currently active filter is automatically reset.
If you want to remove a filter completely, click on the x or – button on the right. The Project Explorer in the navigation area offers a convenient way to do so.