guest lecture

Question 1

how is high quality data obtained? sampig and data quality assurance

Answer 1

selection, invitations and email notifications (sampling)

Assignment, allocation and resampling

Interviewing/data collection: questionnaires
survey management tool

Question 2

Selection and invitation (turbosampling) how does it work

Answer 2

regular sampling runs check the sample needs across each live survey

A proportion of eligible panelists are pulled at random

A panelist is invited if they “satisfy” the requirements

Turbosampling is triggering the invitation batch

A panelist is only allocated to a particular survey once they click on the invitation

Question 3

Getting allocated (sampled and resampled)

Answer 3

1) panelist clicks invite

2) panelist is sent to surrvey management tool

3) survey platfor allocates panelist to a survey

4) panelist take survey

either panelist is resampled

or panelist returns to account page

Question 4

Quality assurance system levels: out suite of globally proven quality processes will help ensure against fraud protection and poor quality or dishonest responses: insurvey level

Answer 4

quality trap question, identifier QC, attention QC and speeding

Question 5

Quality assurance system levels: out suite of globally proven quality processes will help ensure against fraud protection and poor quality or dishonest responses: Welcome survey level

Answer 5

double opt-in, email domain check, country of residence vs ip, age consistency trap

Question 6

Quality assurance system levels: out suite of globally proven quality processes will help ensure against fraud protection and poor quality or dishonest responses: system level

Answer 6

out of towners, suspect panelists, referrer checks, reports

Question 7

Quality assurance system levels: out suite of globally proven quality processes will help ensure against fraud protection and poor quality or dishonest responses: reward fullfilment level

Answer 7

we re-check the QC status for acounts that have placed points redemption requests before we fulfill them

We chck the reward delivered details upon placing the points redemption request against what the accounts profile data claimed

Question 8

Attention checks

Answer 8

attention checks are useful but imperfect method for improving survey quality

someone answering randomly will sometimes give the response (a false negative )

someone might not be engaged and therefore miss the instructions selecting the wrong answer (a false positive)

therefore yougove uses a wide variety of measures to ensure the representativeness, integrity, and accuracy of data

Hundreds of (quality) predictors are used in a machine learning classifier to create a response quality score