Basic Custom Surveys
Acentric’s Basic Custom Surveys include: questionnaire design and programming in Acentric’s own survey tool, questionnaire hosting, provision of survey participants (panelists) where needed, sample weighting calculations for representation and basic analysis and reporting (delivered in a PowerPoint file).
Based on the questionnaire text you provide, a questionnaire is programmed into an online survey tool on your behalf. Some of the features of this process, that are often not available on popular platforms are described below.
Advanced questionnaire features
Masking selectively masks certain items on lists that were not selected previously (or vice versa), thus saving time and improving accuracy. A common example is a brand awareness question followed by a preference question. For instance if a participant indicates they are only aware of certain brands, but not others, then it makes sense to only show them brands they are aware of on subsequent questions.
The order of answer options can be rotated at random in order to reduce order bias where necessary.
Skips & triggers
Skips save time by skipping over irrelevant questions. Triggers look back at previous answers to determine actions, such showing or hiding questions.
Sample weighting for representation
Sometimes surveys are not representative of the target population of interest. If the distributions of demographic variables are known (e.g. from the census) this can be used to calculate weights that adjust for multiple demographic dimensions. When the weights are applied during the analysis any deviations on the demographic variables will disappear or be greatly diminished. More importantly other questions which are correlated with the demographic variables will be more correctly distributed. A good example comes from the world of politics. If one political candidate is preferred by older people and your sample under-represents older people, the preference results may be understated for this candidate. Sample weighting would solve this issue by up-weighting older respondents and down-weighting younger respondents.
Basic analysis may be delivered as an Excel spreadsheet or for an additional fee as a PowerPoint slide deck.
Graphs are created to summarize data at the overall level. They usually contain percentages or averages (means). Some graphs may be ranked, while others may not be ranked if the original order is important to maintain (e.g. age brackets). Once the data becomes too detailed it may be displayed as a table.
Interpretation: The category by the largest percentage of holidaymakers is software for a laptop/PC, followed by camping/hiking products and TVs or projectors. Approximately 9% have never bought any durable product on holiday.
Cross tabulations and banner tables split the results by demographics or other variables thought to relate to the question of interest. Banner tables differ from cross tabulations in that they contain two or more split variables across the top, while cross tabulations only contain one. Statistically significant differences between demographic brackets / categories are highlighted.
Banner table example
Interpretation: PayPal is used significantly more by 15-29 year olds than 30-49 year olds. Cheques are used by significantly more 15-29 year olds than those in the 50+ category. Males are significantly more likely to use cheques and Payoneer than females. Higher income brackets prefer PayPal, wire and cheques; while the lowest income bracket prefers Payoneer. Wire and Payoneer are used significantly more by those in city / suburbs than town / rural. Those who have education up to high school significantly more likely to use Payoneer, while most other payment methods preferred by those with a degree / diploma.