Glossary

This glossary lists words and terms that are frequently used in SoSci Survey and the manual. Click on the list on the right to go to a particular entry directly.

Responses collected in a questionnaire are saved in the data record. The data record contains a case (a row in the tabular display) for each participant (i.e. for every questionnaire filled out). Each case contains multiple variables (columns in the tabular display).

Note: From time to time, the data record term is also used to refer to an individual case (e.g. „Data record 123 contains invalid data“).

Broadly speaking, display resolution refers to how much information needs space on the screen. Most web pages and SoSci Survey's default layouts use a fixed width (measured in pixels). If the screen is smaller than this size, the participant has to scroll vertically.

Screen resolution is measured in pixels. According to statistics from w3schools.com¹⁾, 98% if all users have a display with a resolution of 1024×768 pixels or larger. According to the browser used, part of this resolution can, however, be used for side panels (e.g. favorites, open tabs etc).

SoSci Survey's default layouts are laid out on a screen width of 800 pixels (for a minimum of 99% of users). This means that the eye does not have to wander too far.

If the layout does not offer enough space for complicated questions, the width can be adjusted with a few clicks in Questionnaire Layout (Questionnaire Layouts).

Drag and drop means nothing more than dragging something with the mouse to a certain position and leaving it there. Drag & drop is used in SoSci Survey, for example, to drag questionnaire elements into the questionnaire.

• Click on the element with the mouse and hold the mouse the button over it
• Drag the element to the desired position by moving the mouse
• Let go of the mouse button

The item concept is used in the context of scales and indexes/indices. A scale measures a construct and therefore does not use an individual question to do so: instead, an entire question battery is used. Individual questions in the question battery are referred to as items. Items do not necessarily have to be formulated as independent questions; statements are often presented as items where the respondent should indicate to what extent they agree with the statement (e.g. in a Likert scale).

The scale value is subsequently determined from the responses to individual questions. Often, this is simply achieved by generating a mean value. Items should illustrate all aspects of the construct. In statistical terms, measurement is improved with an increased number of items (a higher level of reliability), because measurement errors are averaged. In practical terms, a meaningful balance between reliable measurement and overloading the participant has to be found.

• Psychological tests measure constructs (e.g. intelligence, aggression, extroversion) using question batteries, that contain 50 items or more. Of course, this is only possible if the participant brings a corresponding level of motivation. Batteries in online surveys with 10 items can already be too lengthy. Now and then, 3-6 items are able to provide an acceptable measurement.
• Cronbach's alpha is used as an estimate of the reliability of a scale. This illustrates the consistency of a scale – i.e. it describes whether items measure the same thing. Paraphrasing the same statement deliver a high value for Cronbach's alpha – unfortunately, some researchers forget that a good scale also has to fulfil other criteria and that sometimes paraphrasing only seems to improve the reliability and puts respondents off.

With scales, items in SoSci Survey are synonymous with the common understanding of items. They are, for instance, sub-questions or statements in a question battery.

However, SoSci Survey misuses the concept for other components in a question. For example, in a selection question, options are sometimes referred to as items. In an open text input, the individual input fields (incl. labels) are referred to as items.

Therefore, an item in SoSci Survey is (at least technically speaking) the next smallest unit in a question. Although this is a little confusing, it actually enables certain features to be implemented:

• You can control, fixed or dynamically, which items (options, sub-questions etc.) in a question should actually be displayed in the questionnaire. This means you can ask the respondent which items (e.g. products, media, brands) they know first of all, and then only have items that were selected previously rated in a scale.
• Questions can be copied and then the question type can be changed. This means a list of items (e.g. products, media, brands) can be created and the same list used in a multiple choice selection (which items are recognized by the respondent), and then have these rated in a scale (how good the respondent finds the items), and the favorite can then be subsequently chosen in a simple selection.

JavaScript is a programming language that supports all modern browsers. When something moves on a web page (open a menu) or changes (e.g. show new emails) without the page being completely refreshed, there is a high probability that JavaScript is being used.

As a general rule, internet uses have the option to disable JavaScript in their browsers. In certain applications (e.g. pure text-based browsers) there are still browsers that do not support JavaScript. Up-to-date statistics regarding the number of people using the internet without JavsScript are not currently available - in 2010 it was less than 2%.

According to statistics from W3Techs²⁾, in January 2012 92% of web pages still use JavaScript. The probability of a user therefore disabling this browser function is therefore relatively low.

Note: JavaScript and Java are two different programming languages for different applications.

See also JavaScript in the Questionnaire

The program SoSci Survey is continuously updated. In the process, the presentation of questions and, in rare cases, their function are occasionally optimized.

Of course, the presentation/function of a question must not change during the survey. Therefore, SoSci Survey uses a compatibility representation for questions that were saved in a previous version. This means: The question is still displayed as it was in the older program version.

As soon as a question is saved again (), the representation of the most current program version is used, regardless of whether changes have been made in the question. Even after importing from another survey project, questions always use the most up-to-date representation.

Note: DSoSci Survey's compatibility display has nothing to do with Internet Explorer's compatibility mode.

The list of questions refers to all the questions (incl. response options) in a survey project.

• In SoSci Survey the questionnaires are ordered thematically in (sections). Every section contains questions and most questions contain items.
• While questions should be organized with regard to their content, the positioning should be considered methodically (e.g. an icebreaker question at the beginning, sensitive questions more towards the end)
• Individual questions are spread across individual pages in a questionnaire in SoSci Survey with Compose Questionnaire. In doing so, the order in which questions are presented to the participant are determined.

In Compose Questionnaire, an ID can be assigned to individual pages (not to be confused with the Comment, which is only there to help you orient yourself).

This ID is used in order to jump directly to a page using setNextPage(), to insert the page into the page order using setPageOrder(), or to determine the dwell time for the page using caseTime().

A panel is a list of people taking part in a study.

Panel studies usually refer to the same people being questioned or examined several times.

Panel providers maintain a large list of people and put these people (usually a proportion of these) forward to take part in surveys. Therefore, there are consumer panels, politician panels etc.

For example, as part of SoSci Survey, the non-commerical SoSci Panel has been formed.

When information is displayed on a screen, the pixel size indication (abbreviated to px) cannot be avoided. A pixel is a point on the screen.

The size of a pixel depends on the display resolution (measured in pixels per inch, ppi for short). Pixels are traditionally smaller on Apple devices compared to on PCs – however, newer laptops or smartphones sometimes have extremely high-definition screens. In this respect, you can never be sure when designing the questionnaire as to how big an image or a scale will actually be displayed on the screen (apart from the zoom function on more recent browsers).

As the screen display is usually orientated towards pixels, the use of other size indicators (e.g. centimeters) does not really make sense for websites. If you want to make an image smaller, only pay attention to the size in pixels.

In SoSci Survey, questionnaire elements are all of the elements that can be used in the questionnaire: questions, text elements, images, audio files as well as PHP code and HTML code elements.

If a psychological construct is being measured with the help of several items (as with a scale), the responses to individual items are normally settled in the evaluation to one amount - or the unit value index; also known as scale index.

More about this can be found in: Scale Indices

The purpose of sections is to order (partly numerous) questions in the List of Questions in a thematic way.

• A section does not appear in the questionnaire.
• The order of question in the questionnaire is independent of sections.
• Each section can contain a maximum of 99 questions.
• Each section had a two-figure abbreviation (section ID). This abbreviation is adopted into the ID of a question. For example, the first question in section AB will have the question ID AB01.

In a SoSci Survey account, one or multiple survey projects can be created. A survey project contains everything from the List of Questions in the questionnaires to collected data.

• Normally, different survey projects are completely independent of each other.
• A survey project can be shared between users on the same servers (Shared Access to Projects)
• It does not normally make sense to create multiple survey projects when collecting scientific data.
  • Exception: completely different variables have to be collected for different groups of people.
  • If you want to collect different variables (e.g. at different times) from the same person, this is more sophisticated when done in the same survey project.

A list of responses for each participant (case) is found in the data record. Each response is assigned a variable so that responses can be compared with each other. An example of these variables would be the age of the participant.

These variables are also referred to as Data Variables in SoSci Survey.

The different data variables and their names can be seen in the Variables Overview.

The internal variables question type can be selected for questions in SoSci Survey. This question type is always required if a value is not asked directly and is determined elsewhere instead and has to be saved in the data record nevertheless - for example, a random number drawn.

At least one item (one variable) has to be created in an internal variable question type in order to save a value. The question will not create variables by itself.

You cannot fail to come across variables when programming. To some extent, variables are containers that can save a value (e.g. a number or text). Usually, the name of a variable can be chosen at will - in PHP, variables always begin with a dollar symbol ($), followed by a lower-case letter (e.g. $randomNumber).

These variables are also referred to as PHP Variables in SoSci Survey.

Please also refer to Variables in SoSci Survey.