Previous research has demonstrated that respondents recognize the formal and graphical properties of a questionnaire. Among other aspect, the shape, proximity and color of response boxes have been identified as relevant characteristics (Couper et al., 2004; Christian et al.; 2007; Smyth et al., 2006).

With respect to open-ended text questions it was demonstrated that the size and the design of an input field to an open-ended text questions have a significant effect on the amount of information reported by Web survey respondents (Smyth et al., 2007). Larger input boxes evoke more information, also a segmented box (with lines) yields more input than a simple box. Based on these findings it has been theorized that respondents interpret the size of the response field as auxiliary information when searching for an answer to a survey question and when formatting the response. Thus, in addition to the question wording the formal characteristics of the response box are also of importance for the amount of information provided by the respondents.

Thus, it has been recommended that the size of the response field should reflect the desired amount of information. If a longer elaboration on a particular topic is expected from respondents a small input fields is less appropriate. However, if the response field is too large it may restrain respondents from answering since the large input box is perceived as a signal that a lot of details and text is expected which might overstrain them (resulting in item nonresponse). Thus, the optimal size of an input box to an open-ended text question is a trade-off of item non-response and the amount of information provided by those respondents who are actually answering the question.