Collect Contact Data Separately

Online surveys often offer participants a lottery. Alternatively, you may want to send interested participants their individual results. In either case, you usually need participants' email addresses.

This would not be a problem using a simple text input - if participants' anonymity were not compromised by doing so. In addition, collected data now become “personal data”, which is often subject to strict rules of data protection.

To warrant anonymity and collect an email address at the same time, SoSciSurvey offers a special question type: “Collect Email Addresses Separately”. Using it is very simple: create this type of question in your project and insert it into your questionnaire at an appropriate position.

The input form offers 2 options, which you can edit and/or hide separately:

• Particpation in the competition and
• Interest in the study

The first two functions store the email address within the project, separate from the remaining survey data.

Important: The input field for the email address is initially hidden and only displayed once one of the options is selected. This is to emphasise the voluntary completion of the question.

Important: The contact data is also Personal data in itself, although contact data is usually less sensitive than the information in the questionnaire. Make sure that you sign the online AVV in the Project Settings on the Data Protection tab before you collect the data.

Note: This question type still causes some problems in multi-language surveys. In order to adapt the response alternatives for a specific language, you first need to switch to the according language (via Further languages) and then edit the question. A simultaneous amendment for all languages is not possible, yet.

A couple of things to consider when storing contact data:

• In general
  • The information cannot be linked to the survey data afterwards. Nor with the questionnaire that the person has filled out. And it is also no longer possible to find out whether the person has filled out the questionnaire in a meaningful way.
  • An e-mail address (or other information) can no longer be changed: Once saved, the e-mail address is no longer linked to the specific interview in order to maintain anonymity, so it cannot be changed. If an address has been saved, only a message is displayed instead of the form. This also means that only one e-mail address is saved per participation.
  • A separate list is created for each question. Each e-mail address or contact can only be stored once in this list (i.e. per question). If an e-mail address is entered again later, only the selected options will be updated.
  • The identification of the question used and the language of the questionnaire are stored with the contact data.
• Type of contact data: Store e-mail addresses
  • Only formally correct email addresses (using an @ and a full stop) are stored. However, the validity of the email address is not checked.
• Type of contact data: Save any data
  • If no value is entered for the height of the input field, only a single-line input field is available, which allows a maximum of 255 characters.
  • If a height (e.g. 100 pixels) is entered, a multiline input field is available. Up to 4000 characters can be entered there. Longer entries are shortened.

Below Download Data you can find a sub-menu Separately collected email addresses. Here, you can download the stored email addresses as a CSV file. This file can be opened with OpenOffice or Excel.

The file has 3 columns: the participant's information (e.g. email address) and two columns indicating whether the options “Lottery” and “Project Information” were selected.

This question is not limited to storing email addresses. Indeed, any contact details, e.g. a postal address can be saved. The information the question expects can be set in Type of contact data. If the Store arbitrary information is selected, the Text input height can be specified. Enter a value here to allow information to be stored in multiple lines (in the default layout this corresponds to 17 pixels for a line of text).

At the moment, this question does not allow multiple pieces of information to be requested in separate input fields. However, a multi-line label can be put before a multi-line input field. Enter, for example, 68 pixels as the height for the input field (in the default layout this is 4 text lines) and for the Text input's label:

  First name:<br>Surname:<br>Street:<br>Postcode, Place:

A check to see whether multiple lines were entered and what was entered there does not take place. However, this is usually not necessary as correct information is in the interest of the participant.

Technically speaking, the contact data that is submitted is stored in another database table as the interview's other data. The only thing that is stored for the interview is whether or not the participant submitted contact data, so that multiple addresses cannot be entered into the lottery by using the back button. Whether or not contact data was entered is only stored internally – this information will be deleted once the interview is completed and cannot be accessed by the project administrator.

Contact data is stored without using consecutive numbers and, according to specification, the data of a SQL table does not have a definite order. However, it is also possible that data is returned when accessed unsorted in the order of storage. Therefore, SoSci Survey returns contact data in alphabetical order.

Collecting contact data separately is a feature for the project administrator, in order to collect personal data separately from other data. In the interest of the project administrator to leave the rest of the data anonymous, they have to forego the following possibilities to bypass anonymity:

• Using another question type (e.g. open text input) to store contact data.
• Accessing contact data at short intervals, in order to track the time of entry by observing the change.
• Accessing contact data before a sufficient number of interviews (e.g. 50) have been stored.

If the feature is not used selectively, and the project administrator does not have direct access to the database, the anonymity of the interview data is assured. If this is not regulated in the terms and conditions, it must be agreed contractually with server operator that raw/crude data is not passed on, which could allow at least one probabilistic allocation based on the sequence.