A survey is a tool for collecting specific information about a fact or product. It can be administrated online, on paper, by phone, email or face-to-face. A survey should consist of two key aspects: a target population to be surveyed and a data collection method.
A survey audience can be known or anonymous. It can also be geographically dispersed elsewhere or in the same location. A Web-based survey normally addresses an audience that cannot be reached by classical means such as face-to-face, by phone or snail mail.
A survey designer may take into account real and imagined users who use products to do something in their own environment. In mass communication research a survey audience is composed of voluntary people who attend a certain content or medium.
2. Web survey
According to the literature, a survey passes through four phases, from the initiation to the closure phase:
- define the survey purpose
- design the questionnaire
- administrate the survey
- validate and interpret results.
Some theorists recommend choosing a sample group of users to be surveyed. A clear purpose of a survey helps gather good quantitative and qualitative data, which ensures the survey’s success. A clear purpose also guarantees that the subsequent three stages are properly applied to finalise a good communication product.
A survey questionnaire should be brief, attractive and with a logical order of the questions. It may contain open and closed questions which allow gathering subjective data like unforeseen and insightful ideas about non-palpable issues.
Some of the questions may invite respondents to choose an answer from a given list. This helps measure the information better. When using scales to define frequency of events and facts, terms like “often”, “sometimes” should be avoided. In the case of rating content-based products, items such as “poor”, “average”, “good” and excellent” may be used. For time periods it is recommended to use clear formulations such as “daily”, “approximately once a week” and “approximately once a month”. Such options may be incorporated into a drop down menu.
It is said that survey respondents tend to have a predisposition towards some prestige bias, which means that they may answer in a way that makes them feel better. They may not lie directly, but may try to put a better light on themselves.
According to MacNealy (1999, p.157) a survey design should take into account three basic aspects: layout, type and scale of questions. Some other researchers recommend a triangulation technique which implies collecting information by using more than one method to increase the survey credibility.
If this is not possible, a solution to achieve similar results, in a Web-based survey, is to tabulate the answers. This means that the results from two or more questions may be compared and interpreted according to a cross-tabulation method. This will increase both researcher performances and user satisfaction.
Some survey questions may require answering a bipolar question (yes/no) or a reflective answer (a question followed by “Please explain”). This way a researcher may ensure gathering qualitative data which may have a significant contribution to the survey results and interpretation. Such reflective answers may give a clear understanding of backgrounds or about factors that lead to a certain position.
A Web-based survey may reach a large number of users, if the audience is not sampled. Researchers believe that it is more efficient than an email-based survey, as some users may consider an email survey as an unsolicited email, and therefore may reject it.
A Web-based survey
- may be cheaper, since classical materials, such as paper, stamps and phone cost, are not involved
- and may be taken simultaneously by lots of users covering various geographical regions.
3. Web surveys: Their benefits and limitations
The amount of data resulting from a Web-based survey might be huge and therefore difficult to handle and process. Moreover, when surveying an international audience, the questionnaire should be ideally available in the native language of the respondents otherwise it may limit respondents’ access and therefore alter the survey conclusions.
From another point of view, as the survey is available on the Web, such a method may ignore users who do not access the Internet. In terms of sample size, a researcher may risk damaging the final results, if the sample is not selected properly. This may generate significant differences that may lead to inconsequential or meaningless interpretations. But this is the case when communication products are distributed through several channels and the Web is one of them.
A Web-based survey remains a powerful instrument to harvest qualitative and quantitative data such as views, motifs, assessments, suggestions, beliefs, actions and attitudes of a Web-based audience in relation to a product. Despite some minor limitations, a Web-based survey allows to investigate needs, opinions and expectations of a large group of people who respond from a distance. Such data cannot be easily collected by other inexpensive means, as users are located at distance.
When considering an international audience, a survey, as every communication product, may mirror various cultural contexts. The task of any communicator to balance cultural assumptions and communication requirements might be challenging.
Burnett R.E. (2003) Technical Communication, Harcourt Publishers
Coe M. (1996) Human Factors for Technical Communication, Wiley
Hackos J. and Redish J. (1998) User and Task analysis for Interface Design, Wiley
MacNealy M. S. ((1999) Strategies for Empirical research in Writing, Allyn and Bacon
Schriver K. (1997) Dynamics in Document Design: Creating Text for Readers, Wiley