Understanding the protocols for questionnaire based data

Priyatam Dutta

I am currently investigating data on phenotypic traits within the UK Biobank dataset, and I have observed a notable variation in participant responses across different fields. For instance, in Category 100027: Fluid Intelligence/Reasoning, 165,349 participants completed the first test (Field 4935: FI1 – Numeric Addition Test in instance 0). However, the number of participants progressively declines across subsequent questions, with only 4,346 participants completing Field 5866: FI13 – Subset Inclusion Logic.

To better understand the reasons behind this missing data, I have the following questions:

  1. Were participants given the option to skip questions in the questionnaire?
    Whether participants could skip questions without being forced to respond.
  2. What factors contribute to the variation in participant numbers for longitudinal cognitive measures?
    For instance, is the decrease in participant numbers over time due to specific attrition reasons, such as participants dropping out of the study, not being shown subsequent questions, participants deciding to not go for the assessment, or other factors?

Clarifying these points will help determine whether the missing data arises from participant behavior or procedural elements of the study design.

Comments

3 comments

Date Votes
  • Comment author
    Melissa L The helpers that keep the community running smoothly. UKB Community team Data Analyst
    • Edited

    Hi Priyatam

    Cognitive function was initially assessed at baseline. The whole cohort attended the initial assessment centre visit. Subsets of the whole cohort were invited to further assessment centre visits (Repeat of baseline, Imaging visit, First Repeat of imaging) where cognitive function assessments where repeated. From 2019, cognitive function was also assessed via an online questionnaire. Over time, there have been changes to the set of cognitive function tests administered. For example Trail making was introduced in 2016. This information is available on  UK Biobank's Showcase. 

    Participants could not skip an answer but could  chose the option 'Prefer not to answer.' In addition, for the example of the Fluid Intelligence test, particpants had a time limit of 2 minutes to answers as many questions as possible. Resource 100231 details the procedure for the Fluid intelligence test. Resources are available for all cognitive function assessments under the resource tab for the relevant field on UK Biobank's Showcase.

    Thank you for getting in touch, we hope this has been helpful,

    Melissa 

     

    0
  • Comment author
    Priyatam Dutta

    Hi Melissa,

    Thanks for the clarification.

    So the reason why there is a steady decline in the participants answering the fluid intelligence questions is due to elapsed time?

    Best regards,
    Priyatam Dutta

    0
  • Comment author
    Melissa L The helpers that keep the community running smoothly. UKB Community team Data Analyst

    Hi Priyatam

    Yes that's correct. You may find data-field 20128 helpful.  It is the number of fluid intelligence questions attempted within time limit,

    Thanks for using the forum, 

    Best wishes 
    Melissa 

    0

Please sign in to leave a comment.