METHODES STATISTIQUES EN GESTION NIELSEN

Transcription

METHODES STATISTIQUES EN GESTION NIELSEN
__________________________Groupe HEC__________________________
METHODES STATISTIQUES EN GESTION
NIELSEN
Adapté de :
J.T. McClave & P.G. Benson
Statistics for Business and Economics
Maxwell Mac Millan International Editions, 1991
___________________________________________________________________________
Ce document ne peut être utilisé, reproduit ou cédé sans l’autorisation du Groupe HEC
NIELSEN
Although the cable television industry could provide viewers with 100 or more channels, a
study by the A. C. Nielsen Co. suggests that that may be many more than viewers want or will
ever watch. The Nielsen survey indicates that as the number of television channels increases,
the percentage of channels viewed for 10 minutes a week or more declines (Landro and
Mayer, 1982). In a similar study, 20 households were sampled, and the number of channels
available to each household was recorded. In addition, each household was asked to monitor
its television viewing for 1 week and report the number of channels watched for 10 minutes or
more. The results appear in the accompanying table.
Household
1
2
3
4
5
6
7
8
9
10
1.
2.
Number of
chanels
available
12
29
4
20
40
5
6
4
14
20
Number of
channels watched
for 10 minutes
6
10
3
8
12
3
5
4
8
6
Household
11
12
13
14
15
16
17
18
19
20
Number of
chanels
available
25
8
5
10
16
4
5
45
35
50
Number of
channels watched
for 10 minutes
10
6
4
4
9
4
1
13
5
10
Make a first study with the descriptive statistics of each variable.
Make a simple regression (with the graph) to explain Number of channels watched for 10
minutes by Number of chanels available.
3.
4.
The scatter is suggesting a « power » model Y = aXb, which is also Log(Y) = Log(a) +
b×Log(X, where Log is the decimal logarithm.
a) Compute the variables Log(Y) and Log(X) by using TRANSFORM/COMPUTE
and select the function LG10.
b) Make the new simple regression.
Give a comprehensive model to your boss, and find the “famous” Nielsen formula :
1