Section 3 :
Base data
We are interested in this research to study a data base
composed by the partially exporting companies with foreign participation
established in Tunisia.
This list is provided by two bases : on the one hand with
the meadows of the Agency of Promotion of Industry (API), on the other hand
with close to the official Web site the API one
The list provided by the API one comprises a set of data on
the companies with knowing :
it corporate name ;
- The address ;
- Telephone number, fax ;
- Activity, Branch of industry ;
- Person in charge ;
- A Number of employment ;
it rate of foreign participation ;
it capital ;
Missing information which is provided by the site of Internet
the API one is :
them nationalities of the foreign participants ;
it creation date of the company
We will try to check our assumptions by the analysis of these
data.
We will try to explain, on the one hand, the relation between
the size of the company and the choice of the mode of presence by the variables
capital and a number of employees and on the other hand, the influence of the
experiment of the company on the choice of the mode of presence by the variable
year of presence.
The variable nationality facilitates the interpretation of the
countries present to us on the Tunisian territory.
Section 4 :
Statistical tools
The techniques of analysis of data that we propose to use are
the descriptive method, the tri dish, the cross sorting and the test of Khi
2.
1 it descriptive method
The principal stage of the descriptive method is that of the
statistical observation, it consists with a transcription of the part of the
reality in which one is interested in a number of data to which the analysis
will relate.
Several approaches are possible ;
seek it of a synthesis by the image which means translation of
the information quantified in the form of graphs ;
seek it of a synthesis by parameters of central tendency and
of dispersion, these parameters are not the same ones if one is interested in a
qualitative character (the average, the median and mode), or if one is
interested in a quantitative character (the interquartile interval, the
variance and the standard deviation).
2 them flat sorting (or simple
sorting)
The simple sorting consists in comparing the answers
concerning a question according to various possible methods', which amounts
revealing the absolute and relative frequencies for each variable on each
method.
3 them cross sorting
The cross sorting aims to cross the variables two to two with
an aim of determining the existence or not significant association between
them, or them and the individuals.
This level it should be noted that the validity of identified
associations was the subject of the statistical tests of significance
(Khi-deux). Thus several crossings can be carried out.
4 it test of Khi-deux
To supplement and validate classifications carried out, one
can plan for the cross sorting the test of Khi-deux, it consists in comparing a
distribution observed with a theoretical distribution.
Indeed, the ÷2 observed is calculated in the following
way :
With :
fij : Frequency observed in the iéme line and the
jéme column
Fij : Theoretical frequency in the iéme line and
the jéme column
m : A number of lines
N : A number of columns
· In our analysis the theoretical ÷2 is in the table
of distribution according to the probability of error á=10% and numbers
it degree of freedom = (M-1) X (n-1). It should be noted that when ÷
² 0>? ² T, there exists a relation between
the two variables. In the contrary case, the two variables in questions are
independent.
| 
