4 Questionnaire Development 4.1 Introduction
4.4 Developing the Online Questionnaire
4.4.1 Design Considerations
Optimal design of web surveys is essential in order to maximise the benefits of such a rich audiovisual and interactive self-administered medium227. However this must be done in a responsible and informed manner227. Since the ever-developing worldwide-web evolves at a fast pace, literature
surrounding the Internet becomes dated very quickly. The Internet offers new possibilities for the implementation of questionnaires. However this also comes with a numbers of issues that included the complexity of survey software, respondents’ computer experience, browsers, speed of access, possibly security issues and complexities with operating systems
configurations228. A review of current literature on developing and designing questionnaires, particularly those that are web based, found that a number of factors regarding questionnaire design influence participant responses229. Factors include, the topic of the survey or study, advanced graphics, instructions for completion and layout229.
Respondents may not actually view the questionnaire as was intended by the designer due to differences between web browsers, operating systems, screen configurations230. This can compromise the aesthetics, particularly images and colours of the questionnaire. A progress indicator is an additional function suggested from user comments in one survey229. Software can place constraints upon the layout and possibly the content of online surveys and they take longer to design than postal surveys, highlighting the importance of correct software selection. HTML surveys enhance response rates because of improved aesthetics with colour usage; screen design and question formatting such as skip logic and conditional questions231.
Concerns have been raised over Matrix style questions as response bias has
been reported232.
Matrix style questions are two-dimensional version of a multiple choice question type with selection boxes or circles. They are arranged in a table- like format with the
questions listed to the left and the answer choices across the top (see figure 4.2). Survey software advise minimal use of matrix questions as they
increase survey complexity with regards to appearance and completion233.
In comparison, drop-down style questions are less complex and are useful when many choices are provided. This question type offers respondents multiple choices in a dropdown list format and requires the selection one response (see figure 4.3).
Literature suggests that it is important to consider technical requirements
such as browser requirements, technology used as mode of access232. In
addition, questionnaires must be trailed thoroughly before they are made
available to participants232. Recommendations have been made to employ
filter questions to prevent participants from having to complete unnecessary
questions232. It is important to ensure respondents reply only once234. A
unique identifier is reportedly the most effect method for preventing or
identifying multiple responses235. Alternative methods would be to compare
host names or Internet protocol (IP) addresses of submissions236. However
Figure 4.2. Example of Matrix Style Question
this requires knowledge of how these addresses are assigned on each network to make effective use of this technique236.
Clear understanding of the target audience is important to design an appropriate questionnaire237. Limited use of images, charts and graphs reduce the questionnaire download time237. However careful use of images can provide a context for questionnaire without having to read extra text237. Advanced graphics may improve respondents’ motivation and satisfaction237. Carefully selection of font and consideration of font size ensures
readability237. Navigation guides such as progress bars are useful tool to help participants complete the survey without getting discouraged or lost237.
4.4.2 Software advantages and limitations
Designing questionnaires using both SS and AFC provided an opportunity to clearly assess advantages and limitations of both survey software. The usability of SS was significantly poorer than AFC. Designers had to edit colour, images, font etc on a separate screen without being able to view immediate results. In addition, only one image may be added, including logos (see figure 4.4). AFC in comparison, enabled users to easily design
questionnaires using a drag and drop editor. In addition questions may be copied and pasted and therefore easily duplicated214. In addition AFC allows a test drive function, which ensures that the questionnaire may be completed smoothly, and data collections function efficiently214. Forms are easy to personalise with any number of logos or pictures214. A useful function is the ease to create a PDF version, create conditional questions and 24-hour easily accessible support. A PDF version is a particularly valuable asset as this can then be printed and sent as a paper copy identical to the online version of the questionnaire. In addition it is possible to collaborate with other users to create forms214. AFC is very customisable, as far as changing
backgrounds, colours, fonts, and adding images are concerned214. The software also enables users to create multiple types of questions and copy and paste multiple form elements214.
Nevertheless, SS software does include basic emailing software, this allows users to control mailing lists but does not include any form of automated email scheduling213. Although answer piping is a feature offered, text must already be included at the start of the survey and this is difficult to set up213. Piping is a technique that involves automatically including a previous answer in the questionnaire as part of a subsequent question. For example if the participant entered their name as ‘Mrs X’, subsequent questions would
automatically include the participants name ‘Mrs X’, making the questionnaire more personal. In addition, a previously selected answer may be ‘piped’ as
new answer choice in a follow-up or future question. A benefit of this type of
survey is that it offers unlimited responses per survey213. Skip logic is
available but only using pages, not individual questions. In addition piping is possible from previous answers but this is complicated to put into practice213.
A limitation of AFC is that there are a restricted number of responses perform. Creating multiple forms can rectify this and all responses can be imported on a single datasheet. In order to link all data for longitudinal analysis a unique ID will be given to each participant to minimise errors that might occur with multiple forms. Although SS allows page condition logic, individual conditional questions are complex to construct and very basic. Question libraries appear to be a useful function but in reality questions need to be designed in the library. AFC allows simply copying and pasting
questions even between different forms. Support for SS is not as useful or available as their website claims but computing staff at the UoL were able to assist with any issues or problems213.
Once the questionnaires had been designed, both survey options were trialled using different browsers such as Internet Explorer, Safari, Firefox and Google Chrome. The appearance of both questionnaires did not differ.
Literature also highlighted the importance of considering participants mode of access such as using a laptop, mobile, tablet or a personal/desktop computer (PC). Many participants may have limited technology to access the Internet so it is important that questionnaires may be completed using as many modes of access as possible.
Time stamps collected by AFC allowed establishment of the exact dates and times at which participants completed their surveys. In combination with MailChimp® , which records times participants open the questionnaire, this data allows the average time to complete of the questionnaire to be
calculated. SS records date of response but not the time completed213. Clearly both the survey software options identified have advantages, and limitations. Is was decided that the most efficient and diplomatic way to evaluate the optimal software would be to use potential participants of the study, the mothers at the Liverpool Women’s Hospital. This would be conducted in the form of feasibility testing.
4.5
Conclusion
Online survey software and email scheduling software are imperative for development, deployment and success of the online LRSQ. MailChimp® has proved to be the most efficient method for emailing participants, particularly with it’s automated scheduling options. SS and AFC were concluded to be the most practical measures to design and carry out the questionnaire
survey. Feasibility testing at the LWH will ascertain the most effective survey software for the LRBCS.