Table 6 This table summarizes the ability of each anonymization method to result in the same analytic conclusion as that of the original data.
From: Comparison of Three Anonymization Tools for a Health Fitness Study
Main policy conclusion | ARX | SDV | SynDiffix |
Children driven by car who live within wheels or walk distance from school should be targeted by interventions promoting active transport | O | X | O |
Descriptive analytics results | ARX | SDV | SynDiffix |
Overall, 43% of the participants reported active commuting modes to and from school and an additional 13% only in one direction | O | X | O |
Participants more often used active commuting from school (56%) than to school (44%) | O | O | O |
Males and females were choosing active commuting equally often | O | O | O |
Females preferred to walk and males preferred to use wheels transport | O | X | O |
The Walk group had the lowest and Public group had the highest median distance from home to school | O | X | O |
The active commuting groups (Walk Wheels) typically live close to school | O | X | O |
CRF-related analytic results | ARX | SDV | SynDiffix |
When commuting from school to home, both the main effect of commuting group and its interaction with distance were significant | O | X | O |
The Car group had a significantly lower predicted value of CRF | O | X | O |
The Car group was the only one where distance of travel was related to CRF having significant difference to the reference (Walk) group; the participants with larger car travel distance had higher predicted VO2max | O | X | ? |
Overall interaction of commuting group with gender was not significant | ? | O | ? |
The largest sample difference between males and females was observed in the reference group (Walk), but it was only significant in the Car group | ? | ? | ? |
When commuting from home to school neither main effect of commuting group, nor it’s interactions with gender and commuting distance were significant | ? | O | O |
Main scientific results | ARX | SDV | SynDiffix |
The main effect of commuting group on CRF and its interaction with distance were significant in the direction from school to home, but not in the opposite direction | ? | X | ? |
Predicted differences in CRF between commuting groups were moderate and generally higher in males than in females | O | X | O |
When comparing commuting group median distance from school to home, males driven by car had around 4 ml/ min/kg lower predicted CRF than those who walked or used wheels commuting (e.g., bicycle, skateboard) | ? | X | ? |
