Based on published collaborative research co-authored by Dr Arash Ardalan.
The research question
Do associations with substance use extend across the distributions of attention-related traits and risky-driving scores?
Study context
The multicentre cross-sectional study involved 1,747 motorcyclists from three Iranian cities. Standard questionnaires collected attention-related traits, risky-driving behaviour and substance-use information.
Statistical approach
- Compare questionnaire scores using mean-based analyses.
- Use quantile regression to examine associations at multiple points of the conditional score distributions.
- Consider results across the two behavioural outcomes alongside the study design.
What the analysis provided
The publication reported associations across multiple quantiles, adding a distributional perspective to the mean-based comparisons.
Cross-sectional and questionnaire-based data limit causal interpretation. Findings need interpretation within the population and questionnaires studied.
Practical lessons for your project
- Match the analysis to the score
- Consider the scale, distribution and interpretation of a questionnaire outcome when selecting a model.
- Look across the distribution
- Use quantile regression when the research question concerns particular conditional percentiles.
- Communicate associations carefully
- Explain estimates and uncertainty without turning an observational association into a causal claim.
Original publication
Maryam Akabri; Yaser Sarikhani; Kiana Khatami; Arash Ardalan; Heidar Safarpour; Mohammad Hadi Imanieh; Tahereh Ostovar; Reza Tabrizi; Arash Mani; Seyed Taghi Heydari; Kamran Bagheri Lankarani
The association between the score of adult attention-deficit/hyperactivity traits and risky driving behaviors with alcohol intake and narcotics consumption among Iranian motorcyclists
Traffic Injury Prevention · 22(3), 189–194 · 2021