So you have been using face-to-face and telephonic surveys for years and you are now wondering if online surveys have any merit? Online panels of pre-recruited, profiled and incentivised consumers provide an effective mechanism for conducting surveys online. Here are a few-key advantages and disadvantages to take into consideration before commissioning your next survey.
Advantages of using online-panels
Data quality is superior since interviewer error is eliminated. The marketing research industry places heavy reliance on freelance interviewers and surpervisors, and a substantial component of quality control is left to these networks.
Online has good penetration in metro areas (31% accessed the internet in the last 4 weeks - AMPs 2011). By covering metro areas you are covering half of all income earned in SA and you are also reaching those with higher average income levels (see graphs below).
Using an online-consumer panel for your survey is more cost effective.
Response rates are generally higher, which means less non-response bias.
It is more engaging for respondents, who have become used to sharing information via the internet rather than face-to-face. Social-desirability bias may also be reduced.
Online is often faster than traditional modes (depending on incidence and questionnaire length).
Ideal for tracking surveys, as panelists are engaged for the long term.
Incentives are commonplace in the world of online panels, while respondents generally go unrewarded with traditional methods.
Allows images and videos to be shown to respondents.
Where online is at a disadvantage, or is at least no better than traditional methods
Males living alone and lower-income people have been regarded as less likely to participate in surveys in general. This was true of face-to-face and telephonic modes, and online is no different.
Internet penetration is still fairly low in the rural areas of South Africa meaning national representation for certain product user groups is not possible. Fortunately, this is often not an issue. In fact, many studies conducted using traditional methods in the past excluded rural areas, either to save costs or because the population of interest resided in urban/metro areas.
Probability samples are more difficult to draw. Quotas or weights are usually required in order to ensure representation.