The Study of Growth - An Invariant Approach to Statistical Analysis of Shapes - page 217

norms or reference data sets (Garza and DeOnis, 1999), sampling
points can be chosen to meet the study criteria. However, most of the
time, sampling cannot be done according to a predetermined design,
and other factors determine the division of this continuum into the
increments that are studied. When growth data are collected from
museum specimens or from patient records, the researcher may have
little control over the intervals of time included in the growth study.
Longitudinal data refer to data points taken from a single individ-
ual at specific intervals over a period of time. Longitudinal data from
numerous individuals can be combined to create samples of longitudi-
nal data. Cross-sectional data refer to data taken at specified ages, but
Figure 5.2 Longitudinal data representing the linear distances between the right and
left foramen lacerum on the cranial base of the New Zealand white rabbit. Note the
variability in the magnitude of growth for each age interval between rabbits. Note
also that the length of this distance decreases with age for some rabbits and some age
intervals. We have found this to be a normal occurrence during craniofacial growth
when studying individual linear distances. An average linear distance was calculated
for each age group and plotted as the cross-sectional mean. The cross-sectional mean
portrays a steady increase in size over time and does not provide information on the
variability in individual growth patterns.