AD ALTA
JOURNAL OF INTERDISCIPLINARY RESEARCH
levels
and the feature B takes on m of the
levels
, whereas
or
. At
first we tested the zero hypothesis
: the features A, B are
independent opposite the alternative hypothesis
: the features
A, B are dependent. As a testing criterium we applied the
statistics
which is defined with the relationship
=
, where
are empirical numerical
data and
are the expected numerical data. The testing
statistics
has, by being valid the tested hypothesis
- a division with the number of degrees of discretion r
=
. We reject the tested hypothesis
at
the level of importance
if the value of the testing criterium
overpasses the critical value
. We will find the
critical value
in the table of critical values
-
divisions.
- with the test of independence for the contingent
table
we will test the zero hypothesis
opposite
Table 3: The relationship between the intergenerational learning and the verbal aggressiveness
the alternative hypothesis
, which expresses in our case that
there exists a statistically important relationship between the
intergenerational learning in the family (between the
grandparents and the grandchildren) and the occurrence of the
aggressive behaviour by the grandchildren.
The test was carried out by means of the programme
STATISTICA. After entering the input data in the output set of
the computer, we obtained the always - contingent table, the
value of the testing criterium
- test and the value p. We can
assess the test also by using the value p, what represents the
probability of the mistake we will make if we reject the tested
hypothesis. If the value of the probability p is sufficiently small
(p < 0,05 , respectively, p < 0,01), we reject the tested hypothesis
about the independence of the observed features A, B (at
the level of importance 0,05 , respectively, 0,01). It means that
the difference between the frequency found in the sample and
the expected frequency is too big to be the consequence only of
the coincidental selection and therefore it is statistically
important. If the value p is equal or higher than the chosen level
of importance, it is not possible to reject the zero hypothesis. It
means that the difference between the observed (empirical) and
expected frequencies can be the consequence of the coincidental
selection and therefore it is not statistically important. In this
way we proceeded in all four cases (more in detail: Tirpáková &
Malá, 2007).
In the following table no. 3 there is shown the result of the chí-
squared test we used in order to verify the hypothesis whereas
we were interested in the relationship between the
intergenerational learning and the verbal expressions of
aggressive behaviour.
Since the value of the probability p is smaller than 0,05 (p =
0,000555), we reject the hypothesis
at the level of
importance
and we accept the alternative
hypothesis
. Based on the mentioned results of the given
test, we can confirm that the intergenerational learning in the
family statistically correlates with the verbal expressions of
aggressiveness by grandchildren. The more frequently the
education between the grandparents and grandchildren takes
place in the family, the smaller is the occurrence of the verbal
aggressiveness by the observed sample of children.
It is evident from the table that by 17,4% of pupils there do not
appear verbal attacks against other people when they have the
possibility of the common activity with their grandfather or
grandmother. Analogically we proceeded by verifying the
hypothesis where we assumed that the intergenerational learning
in the family is related to the physical expressions of
aggressiveness by children. Also in this case we used the
test of independence for the contingent table
- and we
carried it out in the programme STATISTICA.
VERBAL EXPRESSIONS OF AGGRESSIVE BEHAVIOUR
Always
Often
Sometimes
Never
TOGETHER
Learning
n
%
n
%
n
%
n
%
n
%
NO
4
2,3
2
1,1
15
8,7
5
2,9
26
15,1
YES
1
0,6
23
13,4
92
53,5
30
17,4
146
84,9
TOGETHER
5
3
25
14,5
107
62,2
35
20,3
172
100
- 95 -