DATA QUALITY
Data quality has many facets and is always subject to the specific requirements of each individual project. While it is impossible to cover this topic in its entirety, we’d like to present some basic techniques in survey design, that will help to increase the data quality of your project.
Questionnaire Design
Whatever the topic of a survey is, the mental effort required to participate should be kept at a minimum. We recommend to start with general questions and become more specific as the interview progresses. This may require intermediate pages and icebreaker questions to introduce new topics to the respondent. Question texts and answering options should be short, comprehensible and self-explanatory.
User Interface Design
The interface should indicate to the user what kind of input is expected: verbatim text, numerical values or ticked check boxes. Modern techniques in web development, like HTML5 and CSS3, can further support the respondents with providing their data in the right way, e.g. entering the date in a standard format.
Form Validation
In-situ validation of input forms gives the respondent instant feedback about whether the input is in the desired format or not. This feedback can be useful when users are asked to provide information with a standardized format, such as zip codes or telephone numbers or when the numeric input has to sum up to a given total (e.g. 100 percent, 24 hours).
Auto-Coder
If you don’t want to affect the respondent by giving cues, automated coding is worth considering. The user is allowed to enter the data in any format and the survey engine detects the most likely solution in a defined set of options. This technique will allow you to use filters on open text fields.
Questionnaire Logic
Filters and answer piping can help to make your questionnaire logically consistent. Hide brands that are unknown to the respondent in the subsequent evaluation of brand attributes. The same is true for questions that do not apply. In contrast, use answer piping to remind the respondent of what he has responded in earlier questions of the questionnaire.