The Psychology of Eating: From Healthy to Disordered Behavior

178 Obesity

and food diaries (Prentice et al., 1986; Heitmann and Lissner, 1995). In
addition, measuring and comparing food intake of groups of individuals
ignore the variability in the energy requirements of these individuals due
to their body size, composition, and activity levels.

Data analysis

Much of the data are correlational (e.g., associations between changes in
obesity and changes in activity or food intake). This does not enable con-
clusions about causality to be drawn (i.e., does inactivity cause obesity or
obesity cause inactivity?). Further, it is possible that a third factor (as yet
unknown) is responsible for the associations reported. Much of the data
are also cross-sectional (e.g., obese vs. nonobese). This also does not allow
conclusions about causality to be made (i.e., how someone eats when they
are obese may not reflect how they ate to become obese). In addition, some
data are measured at the population level – for example, changes in activity
levels, calorie intakes and fat intakes are all based on measures of popula-
tions. Other data are measured at the level of the individual – for example,
differences between the obese and nonobese. Comparing population data
and individual data is open to the ecological fallacy: Population data can
hide a multitude of individual differences, for the data at the population
level may show one pattern which is completely different to the pattern at
the individual level.

Data interpretation

If the obese do eat proportionally more fat than the nonobese, and this
causes obesity via loss of appetite control, why do the obese not eat more
of everything than the nonobese (i.e., the fat that they consume doesn’t
cause them to feel full, so they then eat more of all types of food)? If
relative increases in fat affect appetite control, why are the population
increases in relative fat consumption not reflected in overall population
increases in food intake (i.e., the more fat we eat relatively, the more we
eat overall)? Finally, if relative increases in fat affect appetite control, which
causes eating more overall, why isn’t eating more overall associated with
an increase in obesity (i.e., as the population eats relatively more fat, it eats
more overall, and gets more obese)? It has been suggested that this reflects
underreporting of calories and accurate reporting of the proportion of fat
and carbohydrate, but this remains unsupported.