Guide to Editing Data
You can learn more about a social network by editing or cleaning the data. Every time you edit data a copy of the original data is set up with the desired changes - the new project title is marked with "(edited)".
Low response rates have a large impact on graphs and subsequently the graph measures.
One way to handle non-respondents (regardless whether they were not asked or did not answer) is to assume their counterparts' ratings of their interaction with the non-respondent.
An example: Person B has not answered the survey, however, Person A has and has rated how often he works with Person B. Using this feature we assume that Person B has the same interaction with Person B.
Occasionally, respondents misunderstand a question or try to inflate their own position within the network. In these cases the researcher can exclude ties that are not mutual and thereby clean the data set.
A mutual tie is when both respondents have identified each other. Setting mutual ties to "Yes", removes all ties that have not been identified by both individuals.
For rating questions, you further have the option to disregard the original rating and set both ratings to the lowest of the two. An example: if Person A rated his interaction with Person B a 5 (e.g., "Very often") but Person B only rated it a 3 (e.g., "Sometimes") then both ratings are changed to 3s.
You might want to analyze a subset of the graph, you do this by removing people from the study.
Manually Regroup People
It is also possible to reassign people to new or existing groups - and then study how these groups interact. This is done manually and you have to set up new groups on the People page first.
Regroup by Answers
Another way to regroup people is by their answers to either scale or multiple choice questions.
For instance, this allows you to see if there's a link between how people interact and their opinions.
If the multiple choice question you asked was limited to one choice, you can group by these answers. You can for instance ask people to identify their department, seniority or location - and then study how these groups interact
Create Group Network
It is also possible to reduce a network to a graph of groups and their interactions. This way groups become the graph's nodes instead of the individuals within those groups.
You define the ties by either "Average number of ties" or "Percentage of possible ties". You will recognize these measures from the group statistics table. Finally, you set the thresholds for which ties to include.
Naturally, you will need two or more groups to use this feature.
Finally, you can manually make any kind of change by exporting the data in Excel format, making the changes in Excel and importing the data again. This allows you to, for instance, add people and ties or to change specific ratings. A guide on how to import data can be found here