Faculty of Public Health - Andalas University - OCS, 13th IEA SEA Meeting and ICPH - SDev

Font Size: 
Digit Bias in Self-reporting of Cigarette per Day (CPD) Frequency: Evidence from Global Adult Tobacco Survey in 28 countries
Ajit Kumar Singh

Last modified: 2018-08-07

Abstract


Background: Self-reported cigarette-per-day (CPD) is an internationally comparable tobacco control monitoring indicator of self-reported tobacco use data. This study estimated and compared digit preference for ‘0’ and ‘5’ using adapted Whipple Index.

Materials and methods: The CPD frequency data for the daily users of manufactured cigarette from 28 country Global Adult Tobacco Survey (GATS) data were analyzed to estimate and compare the adapted original Whipple Index (WI) for CPD and age frequency, digit preference for ‘0’ and ‘5’. Use of mode as an indicator of digit bias was explored.

 

Results: The adapted WI for the self-reported CPD frequency was ranged from 175.6 to 486.5 indicating very rough quality data. The WI for age was ranged between 0 and 289.5. The correlation between WI for CPD and age was negative (r=0.46) but non-significant. The Digit preference for ‘5’ (range: Egypt-2.6%, Panama-24.5%) was significantly lower than digit preference for ‘0’ (range: India14.7%, Egypt-91.2%)  except in India and Uganda. Comparing age and CPD frequency, the concurrence for digit preference for ‘5’ (range: Indonesia-0%, Panama-45.9%) was significantly lower than that of ‘0’ (range:Indonesia-0%, Malayasia-100%). Most preferred mode was 20(64%) followed by 10(29%), 5(4%) and 2(4%).

 

Conclusions : Very rough quality of cigarette frequency in comparison to age frequency is a concern and may be due to pack size of cigarette. Mode can be fair guide for CPD data quality.

 

 

KeyWords(12pt, Times New Roman):  Digit Bias, Whipple Index, GATS, CPD frequency