0851996884.pdf

From these data, the following descriptive
parameters can be computed: (i) n= 30;
(ii)x–= 64.97 eggs; (iii) median = 64.5 eggs;
(iv) range = 70 eggs; (v) variance = 263.48;
(vi)SD= 16.23 eggs; and (vii) SE= 2.96 eggs.
Figure 20.1 shows a histogram presenting
the full distribution.

Describing a Sample with One Trait

Expressed as a Percentage

When the trait is a percentage, several statis-
tical parameters have to be quantified in
order to summarize the main features of the
population from which the sample was
drawn. In this case, the percentage measured
corresponds to the proportion of the obser-
vations of the sample showing a given char-
acteristic (e.g. proportion of individuals that
are females: sex ratio). Four descriptive sta-
tistics have to be computed:

• The total number of individuals for which
  the percentage was determined: N
  
  
  • The number of individuals showing the
    characteristic of interest: x
  
  
  • The percentage itself:
  
  
  • Its standard error: SE
  
  
  
  

In this case, both the percentage and its

standard error have to be provided for an

accurate description of the sample. Here also

a graphic representation can be made and

often a pie chart is enough (see below).

Working example: sex ratio of Trichogramma

In a population of 500 T. brassicae, 163 wasps

were males and 337 were females. So the sex

ratio (% females) was: 337/500 = 0.674 (i.e.

67.4% females), and its standard error was

(i.e. 2.1% females). Figure 20.2 shows the pie

chart corresponding to this data set.

{0 674 1 0 674 500 0 021../.×−( )} =

=

pp×−( )

N

1

p

x

N

=

308 E. Wajnberg

Frequency

7 6 5 4 3 2 1 0

20 25 30 35 40 45 50 55 60 65 70 75 80 85 90

Female’s fecundity

x

Fig. 20.1.Frequency distribution of the data collected to quantify the fecundity of mated T.brassicae
females.