Developing a Quantitative Data Analysis Plan
Page 5 of 12
Example
Screen time and obesity
1. Background
increasing sedentary time, especially screen-time
obesity a mounting public health problem
increasing obesity with increasing screen-time already observed
focus on episodic physical activity but suggestion that total activity more
important
lack of evidence regarding screen-time/obesity relationship in different population
subgroups
2. Aims
to quantify the relationship of screen-time and other sedentary behaviours to
obesity in a large cohort study of older Australian adults overall and within a
range of population subgroups. Particular attention will be paid to how observed
relationships vary with age, disability, work status and lifestyle factors.
Hypotheses to be tested
that individuals with reporting greater daily screen-time will be significantly more
likely to be obese than those with less daily screen-time
that this relationship will not vary significantly according to age, sex, income,
education, region, physical activity levels, work status, disability, smoking,
alcohol, fruit/vegies
particularly interested in how much of screen-
time/obesity relationship explained by physical
activity and variation according to level of physical
activity.
3. Methods
3.1 Data source: 45 and Up Study baseline
questionnaire
3.2 Study population
o definition: Participants in the 45 and Up Study
o inclusion/exclusion criteria: All participants in
the 45 and Up Study, excluding those with
missing data on height, weight and physical
activity
3.3 Study measures
o exposure variables: screen-time, sitting time, standing time, physical activity
o outcome variables: obesity (BMI ≥30kgm
-2
)
o covariates/potential confounding factors: many including age, sex, income,
education, and physical activity
o subgroups to be considered: many
o definitions and derivations
o how missing data will be dealt with: exclusion
o comparison groups: level of physical activity. Work status etc.
3.4 Data cleaning
3.5 Sequence of planned analyses, including
cross-tabulation of relationship of screen-time to potential confounding factors
cross-tabulation of relationship of primary exposures to obesity
RR of obesity (generalised linear models) according to screen-time, sitting time,
standing time, sleep time and physical activity
o age and sex adjusted
While it is ok to include
multiple sub-groups, you
should have a rationale
for doing so, and should
report all results, not just
significant ones