Hi there, I need help to finish this 9-10 pages long (no more than 10 pages), double space, font 12 project. Most importantly, you should be familiar with STATA. For this project, you will be analyzing a provided dataset by using STATA. You will need an introduction to set up a question (attachment 1 has some sample questions). You also have to describe the data, and do some summary statistics as it relates to your question. And then you will need population model, empirical approach, robustness crocus, and finally a conclusion. To finish this project, you will also need to know how to make tables for the data.I will need both the project and the log file of the STATA command at the end please!attachment 1 is the detail explaination of this projectattachment 2 is an example of an empirical projectattachment 3 is the dataset that you will be working on (it shows invalid upload, I can send it to you later via email)please read attachment 1and 2 carefully before you start, thank you!
empirical_project.docx

vasick_honors_thesis.pdf

Unformatted Attachment Preview

Requirements:
1. Use Stata to analyze the provided dataset.
2. Summarize your results in a report. Describe: your question(s), your
data (the specific variables that you use, especially if you have
generated some new variables for the analysis, the subsample you use
if you do not use the full sample using summary statistics and figures),
your regression models, your estimates (in equations/tables), and the
answers to your questions. Remember to interpret your regression
results carefully. Pay attention to both statistical significance and
economic significance.
3. The report shall be no more than 10 pages long (double spaced time
new roman font size 12). Be sure not to include unnecessary Stata
outputs in your report. Use equations and make your own tables
(again, not Stata output tables) to summarize your regression results.
4. Submit your Stata log file together with the report. Make sure that
your results could be replicated by the commands in you log file.
5. Keep in mind that this is a report for an econometrics class. Asking
the right question is important. But to demonstrate that you know
what econometrics models to use and how to properly interpret the
regression results is important as well.
Background of the Stata data file
In November 1984, the Kentucky Public Service Commission (KPSC)
established Administrative Case No. 285 to study the economic feasibility of
providing local2013;measured2013;service telephone rates in Bowling Green and
Louisville, Kentucky. We are only looking at the data in Louisville in this
project.
The experiment took place in the second half of 1986. Before the start date,
Louisville consumers were instructed to choose between unlimited calls at a
cost of $18.70 per month, or a measured service option with a monthly fee
of $14.02. The measured service tariff included a $5 allowance and
distinguished setup, duration, peak periods, and distance. The tariff
differentiated among three periods: peak was from 8 a.m. to 5 p.m. on
weekdays; shoulder was between 5 p.m. to 11 p.m. on weekdays and
Sundays; and off2013;peak was any other time. For distance band A, measured
charges in peak periods were 2 cents for setup and 2 cents per minute. These
charges had a 35% discount in shoulder time and 60% discount in off2013;peak
time. For distance band B, setup charges remained equal but duration
always had a 100% surcharge relative to the corresponding tariff for band A.
The dataset contains randomly sampled households in Louisville. Household
demographic information is surveyed before the experiment took place in
July 1986. Household detailed telephone usage is based on administrative
files recorded by KPSC. In this dataset, we only use household telephone
usage information from Oct. to Dec. i.e., 3 months after the beginning of the
experiment to make sure that households had all been aware of their new
tariff schedules.
Demographic variables
ID
include:
Household ID
MONTH
Month of telephone usage in document
AGE
=1
=2
=3
=4
=5
=6
if
if
if
if
if
if
15-24
25-34
35-44
45-54
55-64
>65
HHSIZE
# in household
ADULTS
# of adults in household
TEENS
# of kids in household
SENIORS
# of seniors in household
EDUC
education of household head
=1 if high school drop out
=2 if high school graduate
=3 if some college
=4 if college graduate
=5 if some graduate school
=6 if graduate school
MARRIED
marital status of household head
=1 if married
=2 if widowed, divorced, separated,
=3 if never married
RACE
race of household head
if white
if black
if spanish
if asian
if other
=1
=2
=3
=4
=5
INCOME
=1
=2
=3
=4
=5
=6
=7
=8
=9
annual household income
if < $5000 if $5000-7499 if $7500-9999 if $10000-14999 if $15000-19999 if $20000-24999 if $25000-34999 if $35000-49999 if > $50000
BENEFITS =1 if receiving benefits such as social security,
food stamps or aid to dependent children
Telephone usage variables
usage pattern variables.
Tariff
include the tariff
variable
anda
bunch of
=1 if measured
=0 if flat
Usage pattern variables are named according to the following system of
rules:
First character: M=minutes
C=number of calls
Second character: A=band A (close)
B=band B (far)
Third character: P= (peak) Weekdays, from 8am to 5pm
S= (shoulder) Weekdays and Sundays, from 5pm to 11pm
O= (off-peak) Weekdays from 11pm to 8am and from Fridays 11pm to
Sundays 5pm
Some Notes (with couple questions for you to start with):
This data file includes a lot of information. You don2019;t need to use all of it.
You could start from pooling all peak calls and minutes, all shoulder calls
and minutes, as well as all off-peaks calls and minutes together and ask
questions like 201C;Did low income people make more of their calls during
offpeak hours?201D;, 201C;Did younger households call more or do elder households
call more?201D;, 201C;Did household call more during the holiday season?201D;, 201C;Did
households who had chosen the flat rate schedule and those who had chosen
the tariff schedule have similar demographic characteristics?201D;, 201C;Did
households under measured service rate make less peak hour calls than those
under flat rate? Did they make more or less distance band B calls during
peak hours? Did they make shorter distance band B calls because of the
duration surcharge? Did these households utilize almost all of the $5
allowance?201D;
If you are willing to program a bit in Stata, you could also recover
household2019;s monthly bill from the information in the dataset. Then you ask
even more interesting questions related to households2019; choice between the
two telephone plans.
Information and Fundraising of Businesses and Charities
By Colin Vasick
A new form of fundraising, called crowdfunding, is rapidly growing in
popularity. Crowdfunding involves presenting an idea to the general public
through the internet in hopes of receiving funding for it. Information
asymmetry could be a problem in crowdfunding because the project starters
have access to much more information about themselves and their projects
than the public has. Websites attempt to facilitate the transfer of information,
but how effective are these methods? This paper uses data collected from
four crowdfunding websites: GoFundMe, Indiegogo, Kickstarter, and Prosper
to study the significance and effects of available information in terms of
convincing the public to contribute to a project. I found that these effects vary
depending on the specific funding model that a website uses. For businesses
seeking donations, information about the project increases funding. In the
case of business seeking loans, information about the project starter is much
more important. Finally, information plays a much smaller role in the case of
crowdfunding charities.
I. Introduction
Crowdfunding is a relatively new form of funding businesses, charities, products,
and other projects. Project starters1 use websites to present their ideas to the general
public in order to raise funds from contributors2. When compared to more traditional
forms of fundraising, such as angel investors, venture capitalists, and banks,
crowdfunding typically raises funds through a large number of relatively small
contributions.
These websites use one of two main systems to facilitate the crowdfunding
process: donation-based and investment-based. The investment-based system is
similar to traditional methods in that the contributors are either providing a loan to a
business or purchasing equity in it. In the donation-based system, contributors do not
receive any financial return, but they may receive perks such as a preorder of a product
being funded, a related gift, or a public acknowledgement. In some cases, contributors
are donating without receiving anything in return.
This paper examines the role of information, both about the project itself and the
project starter, in crowdfunding. I also analyze the differences between information and
its effects on businesses3 and charities4 using different funding models. Information
1
The person who created the crowdfunding campaign and is seeking funding.
Members of the general public who gave money to a crowdfunding campaign.
3
For-profit ventures being crowdfunded.
4
Non-profit ventures being crowdfunded.
2
1
asymmetry is prevalent in crowdfunding as the project starter chooses the information to
present to the public. Most potential contributors only have access to whatever
information is available on the website. This paper examines which of these pieces of
information impact the decisions of contributors, in what way, and to what extent.
The main results indicate that for businesses, a combination of project and
project starter information is associated with the success of crowdfunding. In the case of
investment-based crowdfunding, information about the project starter is greater in
scope, is more easily accessible for potential contributors, and plays a larger role than
in donation-based funding. The role of information in charities is much less significant
than in businesses.
In the remainder of the paper, I begin with a brief discussion of literature relevant
to crowdfunding and information asymmetry and relate the two ideas in Section II.
Section III discusses the data and specific subsets used to answer the question of
information2019;s role in crowdfunding. Section IV estimates empirical models to determine
the effects of information while Section V provides robustness checks and extensions to
these models. Section VI concludes and offers suggestions for future research.
II. Review of Relevant Literature
As crowdfunding is a relatively recent phenomenon, research surrounding this
topic is fairly new. Despite this, there has been some in-depth analysis of this topic on
aspects such as determinants of success of crowdfunding.
Mollick (2012) found that failed projects tend to fail by a lot, and successful
projects tend to barely achieve their goals. This is likely due to people evaluating the
quality of a product, with many people investing in what seems like a high-quality
project. The nature of the sites also allows investors to view how many and how much
people have invested already which further affects a person2019;s decision and prevents
projects perceived as low quality from gaining many funds. He also found that social
networks, the quality of a project2019;s presentation, and the geographic location of a project
starter also play a role in the success of a project. Belleflamme et al. (2012) studied the
influence that pre-orders and other rewards have on funding. Their theoretical approach
found that offering a pre-order of a project is useful in fundraising if the amount of funds
needed is relatively small. Also, special rewards such as gaining exclusive access to the
creation process of a project increases the chances of success as it increases the utility
that an investor gains. However, if a project needs to raise a large amount of money,
pre-orders tend to be less successful. Lambert and Schwienbhacher (2010) looked at
the nature of projects and found that non-profit organizations tended to perform better
than for-profit organizations, and products tended to perform better than services.
This paper differs from the previous literature on crowdfunding since it focuses
on assessing the role of information in crowdfunding empirically. I collected data from
multiple websites for various measures that can provide information to potential
contributors about the project and the project starter. I examined how these measures
relate to the success of crowdfunding. The topic of information that I focus on in this
paper is related to the topic of information asymmetry which is present in crowdfunding.
The project starter has arguably perfect information about himself and his project, and
he chooses what to reveal to the public. Therefore, the classic literature on information
2
asymmetry can shed additional light on the understanding of the role of information in
crowdfunding.
Akerlof (1970) discusses information asymmetry in the used car market. He
discusses how buyers are not aware if they are buying a lemon or a good-quality car.
Used car salesmen have an incentive to sell lemons for higher prices by passing them
off as quality cars. However, with no system in place to deal with information
asymmetry, nobody will be willing to purchase a car for a positive price and the market
collapses. Spence (1973) discusses a method used to prevent this called signaling. He
discusses how the party with more information can attempt to signal the party with less
information. The example used is employees trying to get hired at a company. They use
college education as a signal of how skilled they are at learning. Although college may
not make someone more skilled at learning, one who goes through college can use this
to demonstrate that he is skilled at learning. Those who are less skilled at learning will
have a harder time going through college and be less likely to be able to send this
signal to the employer.
Both of these principles come into play in crowdfunding. Project starters know
much more about their character and the fine details of their projects than anyone else
knows. Websites allow project starters to present information to the public. This
provides direct information in the form of details about the project and the project
starter. There is also a signaling effect as various behaviors can tip off potential
contributors about other information related to the project and project starter.
III. Description of Data
I focus on analyzing information in three types of crowdfunding models. The
dataset used to conduct this analysis was compiled from four websites: GoFundMe,
Indiegogo, Kickstarter, and Prosper. The following sections describe the three models,
the data for each, and the collection of data in detail. Some initial analysis is also
presented and discussed.
A. Donation-Based Crowdfunding of Businesses
To analyze donation-based crowdfunding of businesses, I collected data about
projects from one of the most popular and successful crowdfunding websites,
Kickstarter. This website uses an all or nothing (AoN) system meaning that project
starters only receive money if their goal is met. If their funding goal is not met, the
money is returned to the contributors. I used the website2019;s search functions to find
completed crowdfunding campaigns from the neighboring Californian cities: Davis,
Sacramento, and West Sacramento. I used this constraint in order to limit the variations
that geography has on the success of crowdfunding campaigns5. For each project, I
manually collected data pertaining to the success of the campaign, information about
the project, and information about the project starter.
Table 1 shows summary statistics for the variables about information presented
on the Kickstarter website. These are calculated for all projects and subsets of projects
that successfully reached their funding goals and those that failed to reach their funding
5
As found by Mollick (2012).
3
goals. The final column reports the differences between the means of the variables for
successful and unsuccessful projects with results of t-tests for the significance of these
differences. The variable, # of Facebook Friends, has fewer observations because not
all project starters chose to connect their Facebook accounts to their crowdfunding
campaigns.
Table 1 2013; Summary statistics for donation-based crowdfunded businesses.
Variable
Description Length
All
391.50
(354.40)
Successful
360.66
(293.73)
Unsuccessful
424.10
(408.14)
Difference
-63.44
(52.93)
# of Videos
1.01
(1.11)
0.95
(0.70)
1.07
(1.42)
-0.12
(0.17)
# of Images
1.50
(3.92)
1.39
(3.85)
1.61
(4.01)
-0.22
(0.59)
FAQ
0.15
(0.36)
0.15
(0.36)
0.15
(0.36)
0.0027
(0.054)
# of Updates
3.98
(4.85)
6.23
(5.51)
1.60
(2.32)
4.63***
(0.63)
# of Comments
0.44
(2.16)
0.72
(2.96)
0.14
(0.44)
0.58*
(0.31)
Biography Length
99.26
(92.41)
99.57
(100.51)
98.94
(83.58)
0.62
(13.86)
# of Websites
1.75
(1.45)
1.80
(1.41)
1.70
(1.50)
0.10
(0.22)
# of Facebook
1
Friends
675.18
(755.94)
809.10
(905.00)
526.39
(518.31)
282.71*
(167.15)
# of Contributions
to Others
1.97
(4.19)
1.99
(4.09)
1.94
(4.32)
0.047
(0.63)
# of Past Projects
0.095
(0.36)
0.11
(0.40)
0.080
(0.31)
0.028
(0.054)
# of Successful
Past Projects
0.028
(0.17)
0.043
(0.21)
0.011
(0.11)
0.032
(0.024)
Observations
179
92
87
Notes: Means are reported with standard deviations in parentheses (standard errors for the Difference
column.) The variables: # of Videos, # of Updates, # of Comments, # of Facebook Friends, and # of
Successful Past Projects were tested assuming unequal variances while the other variables were tested
assuming equal variances. Unequal variances were assumed if the ratio of the larger variance to the
smaller variance was greater than 2.
******Significant at the 1 percent level.
******Significant at the 5 percent level.
******Significant at the 10 percent level.
1
n = 76, 40, and 36 for the first three columns, respectively.
Description Length is the word count of the project2019;s written description, and
Biography Length is the word count of the project starter2019;s written description about
himself. FAQ is a dummy variable equal to 1 if the project starter chose to include an
4
FAQ on the project page. # of Comments refers to the comments made by the project
starter in the comment section of the project page which is an area for the public and
project starter to converse and interact. # of Websites represents the number of links to
external websites that the project starter provided. These can be to his social pages,
media pages displaying his works, or his personal websites. # of Contributions to Others
is how many contributions the project starter has made to other people2019;s projects. # of
Past Projects and # of Successful Past Projects describes the project starter2019;s history of
the other projects that he started on the website.
The only difference between means that is significantly different from zero at the
0.05 level (and the 0.01 level as well) is the number of updates that the project starter
made to his crowdfunding webpage throughout the campaign. This is higher for
successful projects suggesting that updates might convince people to contribute and
improve the chance of success. At the 0.10 level, the differences between means for #
of Comments and # of Facebook Friends are significant. Once again, these are both
higher for the successful projects hinting that increases in either of these improve the
chance of success. Overall, about 51.4% of projects are successful. Of projects that
have a connected Facebook account, about 52.6% of projects are successful. This
suggests that the presence of a Facebook account does not improve a project2019;s chance
of success6. This also provides evidence against a sample selection bias when only
looking at projects that were connected to Facebook. This is important in assessing the
empirical models and will be discussed further in Section IV.
B. Investment-Based Crowdfunding of Businesses
The data for investment-based crowdfunding were downloaded from the website,
Prosper, which makes its data available to the public. Information about the city of the
project starter was not available so I was not able to constrain by geography in this
case. However, the available data did contain several variables related to information
and the success of the projects. Prosper uses a loan system in which project starters
pay back contributors with interest. It also uses an AoN system.
Table 2 shows summary statistics for the relevant variables for investment-based
crowdfunding for all projects, successful projects, and unsuccessful projects.
Differences between means and the results of t-tests are presented in the final column.
Homeowner is a dummy variable equal to 1 if the project starter owns a home.
This variable and the Debt to Income Ratio are self-reported by th 2026;
Purchase answer to see full
attachment