Geoscience Reference
In-Depth Information
0.5
0.4
0.3
0.2
0.1
0
20
40
60
80
100
120
Observation Number
FIGURE 11.1
Index plot of mercury concentrations with study year adjacent to the individual points in the
plot.
given year is on the vertical axis, and the observation number is on the hori-
zontal axis. To help display the trends, each data point is coded by the year
of the observation (0 to 11 for years 1995 to 2006). The horizontal line on the
index plots shows the overall average concentration for all 120 observations.
A general observation from the plot is that the majority of the points above
the horizontal line correspond to the later years of the study (i.e., years 6-11),
indicating there was more contamination toward the end of the study than
at the beginning of the study. With this type of plot, it is difficult to explore
the data pattern in more detail than this and to make observations about the
detail of the trend and whether the same trend is observed at each station.
Box plots showing the distribution of the measurement for each of the 12
study years are shown in FigureĀ 11.2. Mercury levels were certainly trending
in a positive direction over time. The spread in the box plots indicates that
the mercury measurements were variable from one station to the next, sug-
gesting that some stations were associated with higher levels of contamina-
tion than others.
Before introducing various methods for quantifying trend, it is often the
case that we would like simply to compare the response across two time
points. If our interest is only in how stations changed between 2 years (e.g., the
initial and final year of the study), a paired t -test can be used. One proviso
is that the data in both groups should be normally distributed, although as
long as the data set is large enough the test is robust for departures from
normality (Sawilowsky and Blair, 1992). The test statistic is
Search WWH ::




Custom Search