GBU 3401

The Effect of Advertising on Elections

Project done by:

Ali Benseddik

Meriem Haddouchi

Wassym Zaz

Spring 2010

Contents

Abstract 2

Introduction 3

Problem Statement: 3

Research Questions 3

The hypothesises: 4

Statistical methods and techniques 4

Data Collection 4

Inferential Statistics: 5

1. The Chi-square test statistic: 5

2. The simple linearregression: 6

Conclusion: 9

Abstract

The purpose of this research is to find out the effect of total advertising costs on the U.S citizens voting. The team members wonder if wining the last elections in the united state is essentially a matter of how much the potential candidate spent on advertising. If our assumption is false, then more important factors would affect the voting ofcitizens and adverting costs would be a secondary factor. In the extreme case would not affect. In fact, during this research the team members aim to see first whether the success of the U.S president in the last elections is dependent or independent on the U.S citizens. Then to conduct a second study to see if the variability of the total advertising costs affects heavily the U.S citizens. Ourresearch is time independent, For that research, the team members decided to collect data of the two main candidates of the last 2008 elections that is Obama and Mccain. In term of data, we gathered the total costs spent by each of the two candidates on each U.S state. The advertising costs that we used are summing up all kind of advertising costs spent during elections. Each advertising cost includesbroadcast advertising, print advertising outdoor advertising, covert advertising and surrogate advertising.

Introduction

Problem Statement:

What is the effect of advertising on elections in the US?

Research Questions

The objectives of the search study are as follow:

• To explore the effects of advertising on the elections results in the US.

We will find out these effects by➢ Examining the degree of exposition to advertising versus the implication in politics.

➢ Comparing the percentage of votes to the amount spent in advertising by each of the 2008 elections candidates and taking into consideration the results of 48 states.

The hypothesises:

Hypothesis 1: The amount of money spent in advertising impact directly the number of votes eachcandidate gets.

Hypothesis 2: There is no relationship between the amount of money spent in advertising and the number of votes each candidate gets.

Statistical methods and techniques

In our research, we use both the descriptive and the inferential statistics. Concerning descriptive statistics, we use this method to conduct a numerical analysis of the obtained results. In additionto the numerical analysis, the graphical evaluation of the data is very important in our research. For the inferential statistics, we mainly use them to draw possible conclusions about the population. In addition to that, we tend to use two different statistical methods, Chi-square and simple linear regression, in order to test our hypotheses. Actually, the purpose behind developing statisticalevidence is to measure the strength of relationships between the independent and dependent variables used in the survey.

Data Collection

We mainly based our research on data collected from the US government elections’ results web site. This enabled us to get concrete and relevant data about the advertising costs of each the parties campaigns. Moreover, we were able to relate the previousdata to the weighted average of votes and that concerning each different state. We believe that this way we will have a good insight on whether or not votes are driven by advertising, and if advertising costs and votes are relatively proportional.

Inferential Statistics:

The Chi-square test statistic:

H0: There is no relationship between advertising expenses and election results....