Release notes Dub Knowledge 2.4.0

• Sample Management
• General
• Feature Management
• QuickList
• Export
• Report/Statistics/Quick Report

Sample Management

Prediction system.

Prediction is a new button added to the 'Sample creation wizard' and its last screen. It is an aid that allow you to estimate the gross amount in sample during sampling (the real number of members that should be invited) to reach the global total 'NET N'. Prediction can help you to calculate the gross values if you set the prediction parameters and enter the total net amount you need. Using the responsiveness during the sampling is some kind of oversampling that helps you to define the total amount of respondents needed as it seems to be necessary and not more. You can measuring and store the responsiveness by country/age(yearly)/gender also using other user defined factors for estimation.

Example:
We have created four cells based on age and gender. From each cell we need certain percentage that you need to fill in to activate the button 'Prediction'. (Without this the button will not be activated and you won't be able to select it):



Press the button 'Prediction' the following screen will open. There are two options on which you can base your prediction:

• Cells of your sample
• Global factors - for example about the survey and other circumstances which might change the responsiveness of the members in the given projects = global things. If you use a factor in a sample then it affects all the cells equally. The factors can be like "subject matter(fun/regular/boring)", "survey length (short/regular/long)", "fieldwork period (normal/weekend)", etc. In case you use more global factors at the same time then their factors will be combined (multiplied). See example below.

Cells of your sample

This example will demonstrate the first option - prediction based on cells 'Use the cell based factors'. Into the field 'Desired Net N' enter the total number of interviews that you need to achieve. In our example we need 1500 completed interviews. When you enter this number the button 'Ok' will activated :



This is the result of the system prediction how many interviews you really need to achieve your target 1500:



The system has calculated that we need sample with 689 members to achieve our Net N target 400. This calculation is based on responsiveness – system feature. See explanation below:

Each panelist has a special system feature 'Responsiveness (percent)' (feature_id=1285) and how this feature is calculated:

It is hard coded into the 'invite _responsiveness' function and into the 'add_questionnaire_feature-value' stored procedure. It works like this:
The system checks the last 10 projects in which the given panel list was involved and count the unsuccessful invitations and calculate the responsiveness percent by the following formula:


(10 - unsuccessful invitation) x 10 = responsiveness percent


In case the respondent has less then 10 total invitations than it is assumed that the missing ones were also completed. This means that if have invite one respondent only 5 times and he/she has participated 3 times it is counted as he/she has participated in 8 from 10 (80%). This means that new members will have in the beginning a very good score.
In case you invite one respondents for example over 50 times then all old projects, except the last 10, will be ignored.

Note that you can also use both factors 'global' and 'cell based'.

Global factors

In order to use this option you should first define any number of global factors and assign as much options you wish per factor.

Define New prediciton factors

As mentioned above you need to activate the button 'Prediction' in the last screen of 'Wizard for creating a new sample' to do so you should define percentage for each cell. After you have done this press button 'Prediction' and check the box 'Use global factors' there are no predefined factors thus this means that you will have an empty list. To define a new factor press button "New" (note that this is only active if you select option 'Use global factors'):



Give you factor a name and description and add options (see our example). An option is a pair of a name and a percent value. This percent value is used (dividing the net) to calculate the gross amount. Later you can edit or delete factor itself or even its options. You can also change option order, use buttons 'Up' and 'Down'. In case you need to define the factor in different languages use button 'Language'.
After you have defined all factors you can check which to use to calculate the prediction, by checking the box in the column 'Use'. You can also change its 'Option', double mouse click will activate a drop down that will show all options that you have defined for this factor. (Note that double click on any other column will open the factor editing scree.)



After you select factors and choose options you can immediately see the combined percent value on the bottom of the page 'Combined global factors' at this moment it is 90.25% . It is the value which is going to be used for the prediction.

In our example we have defined that we want 25% percent from each group for distribution we have used the above described factors and entered 400 into 'Desired Net N'. The result will be following:



The prediction system has calculated the 'Gross N' values per cell. It uses some kind of oversampling method if it's turned on before you enter the prediction screen.

Note that you can recalculate the whole prediction as many times as you want before you draw the actual sample. Further it is important to mentioned that at this moment it is working only in 'cell mode'. Prediction button is not available when you are in 'Rims mode'. However you can use the rims to distribute the percents into cell level and then use the prediction.

Dynamic filter can be used during Sampling.

There is a new option implemented into Sample panel and Scheduled Sample panel. That allows you to load dynamic filter directly into the selection. This function is very similar to known function 'Presets' which allows you to save selection for later use:



This tab option is also available in edit Dynamic Filters panel and new and edit Dynamic Target panel.

Export extra features directly.

• Overwrite existing members in the project
• Add only non existing members to the project and keep them in the sample

New option add features to sample that already exist and has been already imported into DubQuest.

You should create a new sample with those features. In the first screen of the wizard press button 'Export to Dub':

In the next screen select project into which you want to add extra features, use drop down menu. Press button Columns and the system will load all columns that exist within this project. You will be able to edit/match them with features from Dub Knowledge or Add extra feature use button 'Add'. A line will be added at the end of the list. Double click in the filed code allows you to type in the name of the property as it will be displayed in DubQuest sample management (after you have typed the name press enter). You can add as many properties as you need. After this you should match them with features from Dub Knowledge. Double-click on the '-' in entity_id column will opens pop up aplet, which allows you to select and/or search the right feature.:

System info/ Sampling Report (Monthly) an extra event column has been added after the 15th day and at the last column of the tables.

System info/ Slow samples: new filtering mode: 'days'.

General

Customizable user interface.

New menu option has been added to the left menu of Dub Knowledge. You can define your own interface :

• Language - the following languages are available
  • English; Dutch; German; Hungarian
• Look and Feel - you can select from four styles
  • Default(Windows);Liquid; Metal; Motif
• Show button hints – uncheck this box if you do not wish to view help text
• Change password - only for administrators

A new system feature has been added to the tree structure.

This feature allows you to monitor a quality of respondent. The length of four last interviewers is stored here in form of points. In the feature 'Trust Score' is stored average from last four completed interviewers :

How it works: The system will calculate this score for you, read more about this background calculation below. It is important to calculate/fill in this feature with point after the actual project is completed and you have all data in your dataset. In case you calculate this sooner your trust score might not be objective (the data will be calculated from uncompleted results.).
To store a right data in this feature and make the best use of this unique measurement you need to make sure that:

1. You have set up a connection between DubQuest and DubKnowledge:
   
   
   
2. Your project is finished, this means that you have all data from all respondents in the dataset. The project doesn't necessary need to have status closed in ManageFrames.
3. Go to DubQuest and choose Response -> Results -> Put Trust Scores into Dub Knowledge:
   
   
   
   The system will store that score into the 'Last trust score'.

Trust Score calculation

Import metadata from XML.

New option has been added to left menu of Dub Knowledge that allows you to upload metadata in XML format. There is a new option, button 'Export' added to Feature Management screen (see Feature Management) that allows you to export selected feature into XML format. This file can be edited and you can upload it again with above mentioned option 'Import metadata'.

Feature Management

Export to XML

New button has been added to Feature management screen called 'Export'. This allows you to export features from your tree structure into XML format.



What you need to do is select the feature that you wish to export and than click on the button 'Export'. The system will create the xml file and a download pop up will appear:

Advanced feature management menu

New context sensitive menu showing the percentage of feature filling. For enumeration values it is possible to show and directly edit the values in one screen.



Select enumeration feature left mouse button activates the context sensitive menu. Click on the value will lead you to 'Edit enumeration screen':

QuickList

Give open text name to QuickList columns.

Possibility to give open text name to the columns of a QuickList (edit, new, and view):

QuickList columns - combined filtering.

Possibility to to use more than one dynamic filter at the same time:

Export

Export enum and multiple enum.

Possibility to choose a format in which the feature should be stored (if the multiple export mode is categories) as code or text. A new column has been added is active only for enumeration features:

Report/Statistics/Quick Report

Extra flag for reports if they are active or inactive (to refresh or not).

Having a tool to check which Report of QuickReport was not viewed within a time period (system setting) and switch the flag to inactive: