Sampling books the panelists, inviting them to participate in the study
To sample go to the 'Project Monitor'. To sample for the whole target group, use the target group's menu (no.1 in the screenshot below). Sample at target group level would sample in all active quotas, in accordance with their remaining wanted completes, based on the amount of completes you sample for.
To sample in a specific quota, expand the target group by pressing arrow to the left of its name, then use the quota's menu (no.3)
For each sampling, indicate the wanted number of completes and when to send the invitations out.
By default a reminder is set to be sent out 2 days after the initial send out. You have the possibility to not schedule a reminder and change the time period.
Sendouts for closed quotas/target group/projects are removed automatically, so a reminder scheduled after the project closure will not be sent out.
You'll be warned if there are not enough survey links for the sample size you want.
SENDOUT RESULT
The result of a sendout is displayed as a status to you.
Sendout status 'A scheduled sendout is right now actively gathering respondents for your study'
This indicates the system is right now inviting panelists into the study.
Sendout status 'The sendout couldn’t find any respondents. Please try to sample again at a later time.'
This status indicates that there were no panelists available to invite, when the sendout was scheduled to invite them. It does not create a selection since no respondents could be find. You’ll need to sample again, unless the Optimizer got your completes.
Sendout status 'You might need to sample again as your latest sendout is estimated at X survey starts' (no.4 in the screenshot)
This status indicated that at the time of the sendout activation, based on the invited panelists, the system doesn’t estimate that all the needed panelists will answer the study. Which means that you might not get all the completes you sampled for and you might need to sample again.
Here's why:
The new feasibility algorithm works with the number of wanted completes and number of started interviews.
When you create a sendout (sample into a quota) you tell the system how many completes you want, in this example 12.
Based on how many respondents started the survey and how many of those actually completed the survey, the system calculates a conversion rate. So when you schedule a sendout the system will estimate how many respondents need to start the study in order to get those 12 wanted completes. These are estimates.
But when the sendout activates(is sent out) then the feasibility checks which respondents are available at that time and picks the best crowd for your quota that it can find. After selecting the best available at that time, the feasibility algorithm checks each respondent’s answering history and estimates how likely it is they will actually start your survey. If only 10 are estimated to start the survey then you will get the status “Respondents shortage – got 10 out of the 26estimated started”. And if only 10 respondents started then you cannot get the wanted 12 completes. In this case you will have to sample again at a later time.
This is much more accurate than working with invitations. Inviting 100 respondents might yield 80 completes if the selected crowd is made of active panelists, but might yield 5 completes if the panelists aren’t very active. This is why the new algorithm actually look at every respondent’s behavior in the past study invites they got and the system chooses the best available for your study at that point in time.
Sampling tips
- Start with a soft launch to collect about 10% of total quotas. This will help you to understand true IR, LOI and remaining feasibility.
- Make sure the auto close option is on for all sub-quotas.
- Try to utilize the Optimizer as much as possible but after full launch only (you should not use it during soft launch to avoid oversampling).
- Make sure to close all quotas that are filled. This will help you avoid over quotas.
- As soon as fieldwork is closed, reconcile all complete IDs and close your project as early as possible. This will help us to arrange incentives for respondents in a timely manner.
- Why does the system automatically pause projects?
- We have system alerts in place to let the project managers know when the specs of your study is outside of the minimum or maximum rules for each country.
- We have system alerts for IR, LOI, CR, DOR and QF
- By default, low(min) and high(max) values have been assigned for each target group based on company policy. The system will pause sample as soon as the IR, Dropout rate, Conversion Rate or Quota Full values drop below the minimum or go above the maximum set values.
- The rules are different in each country. If you notice during fielding that your target group is no longer approved for send outs, please contact our project management team for assistance.
- The limits for the countries are listed in the chart below: <Insert chart here or location where this information can be found
Comments
0 comments
Please sign in to leave a comment.