BE603 GROUP PROJECT

QUARTO (.QMD) TEMPLATE

1. The case

Crowdfunding platforms like IndieGoGo and Kickstarter have become increasingly popular,
enabling individuals, businesses, and non-profits to secure funding for their projects. The success of
these campaigns, often raising millions, highlights the unique opportunities these platforms offer for
entrepreneurs and innovators. Beyond financial gains, crowdfunding allows creators to test ideas,
receive instant feedback, and build a community or user base.
In this project, we explore the effects of quality and engagement on the amount of funds raised.
Formally, we want to support or reject the following hypotheses:

2. Hypotheses

1. Hypothesis 1: There is a significant relationship between crowdfunding campaign quality

and the total funds collected, through the level of engagement of the backers in the
campaign.
2. Hypothesis 2: There is a significant relationship between crowdfunding campaign quality
and the level of engagement of the backers in the campaign.
3. Hypothesis 3: The relationship between crowdfunding campaign quality and the level of
engagement of the backers in the campaign is moderated by the amount of effort the
campaigners engage to interact with the backers.

3. Assumptions

In this project, we will use OLS regressions models which make the following assumptions:
• Linearity: the relationship between the independent variables and the dependent variable is
assumed to linear
• Normality of errors: The errors should be normally distributed
• Homoscedasticity: the variance of the errors is constant across all the levels of the
independent variables

Stockholm School of Economics. Sveavägen 65. Box 6501. SE-113 83 Stockholm. Sweden. Phone +46 8 736 90 00. www.hhs.se
 • No or little multicollinearity: occurs when independent variables in the regression model
are highly correlated with each other
• No outliers

4. Variables

• When considering which control variables to include, we must assess the quality of the
variables and the logical reason for their inclusion:
– Images and video: These two make up the campaign quality variable and are
therefore excluded due to multicollinearity concerns.
– Business venture and social: These may contribute to higher fundraising by
signaling some form of seriousness. However, when they are both included, their
signficance decrease and so exclude business venture due to the risk of
multicollinearity.
– Creators: the number of creators in the team could impact the overall quality and
management of the campaign. More creators might contribute to diverse skills and
perspectives, influencing campaign outcomes. However, not adequate
– The length of the pitch and the presence of filler words can affect how backers
perceive the campaign. Including these controls helps address the potential impact
of pitch characteristics on engagement and funds raised.

5. Analyses
# Loading the data
df <- rio::import("./data/campaigns.Rds")

# You can use this code chunk, or create new ones, to complete your analyses
# Alternatively, you can create a table for the description of multiple variables.
For example, Montly Income and Age.

vtable::st(
# The dataset to describe
df,

# Vector of variables to display

vars = c("collected_funds", "campaign_quality", "comments_count", "updates_coun
t", "pitch_size", "social", "creators" ),

# How to name the variables in the output table

labels = c("Collected Funds", "Campaign Quality", "Comments Count", "Updates Co
unt", "Pitch Size", "Social Venture", "Creators"),

# Statistics to display
summ = c("mean(x)", "sd(x)", "notNA(x)", "min(x)", "max(x)"),

# How to name the columns in the output table

summ.names = c("Mean", "S.D.", "N", "Min", "Max"),

BE603: Data Analytics III 2 Version: Fall 2023

 # Number of decimals to display
digits = 3,

# Table to export for MS Word

out = "kable"
)

Summary Statistics
Variable Mean S.D. N Min Max
Collected Funds 2572 3079 3000 500 55082
Campaign Quality 5.8 2.69 3000 0 9
Comments Count 15.7 24.5 3000 0 745
Updates Count 3.77 7.56 3000 0 148
Pitch Size 583 382 3000 6 5014
Social Venture 0.288 0.453 3000 0 1
Creators 2.1 2.03 3000 1 43
Looking at the summary statistics, we can conclude that the funds collected, the quality, number of
comments, uppdates and pitch sizes for the 3000 crowdfunding projects within the dataset vary
significantly and may thus contain outliers. Social is a dichotomous variable with a mean below 0,5.
This suggest that most projects are not social.
library(ggplot2)

ggplot(df, aes(x = updates_count)) +

geom_histogram() +
labs(x = "Collected Funds") + theme_classic()
`stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

BE603: Data Analytics III 3 Version: Fall 2023

6. Hypothesis 1

There is a significant relationship between crowdfunding campaign quality and the total funds
collected, through the level of engagement of the backers in the campaign.

6.0.1 Variable Selection

• Dependent variable (Y): collected_funds (total funds collected)
• Independent variable (X): campaign_quality (Composite variable indicating campaign
quality)
• Mediator (M): comments_count (Number of comments by backers, indicating
engagement levels)

In our analysis, we’ve structured our variables based on their roles in the hypothesized relationship.
Collected funds are placed as the dependent variable, aligning with our hypothesis that it’s the
outcome being influenced. Campaign quality takes the role of the independent variable because our
hypothesis suggests its impact on the collected funds. As our hypothesis indicates the influence on
collected funds goes through engagement level which we thus has chosen as mediating variable. To
measure the engagement level, we selected the comment count variable, since it includes the

BE603: Data Analytics III 4 Version: Fall 2023

comments posted by the backers on the campaign page by the end of the campaign. The choice of
this variable also aligns conceptually with the expected impact on funds raised through engagement
within the crowdfunding platform.

6.1 Investigating conditions for mediation

Here we will conduct step one trough three:
Step 1: Regress Y on X.
Step 2: Regress M on X.
Step 3: Regress Y on X and M.
model_campaign_quality <- lm(collected_funds ~ campaign_quality + social + pitch_s
ize + creators,
data = df)

model_engagement <- lm(comments_count ~ campaign_quality + social + pitch_size + c

reators,
data = df)

model_mediation <- lm(collected_funds ~ campaign_quality + comments_count + social

+ pitch_size + creators,
data = df)

summary(model_campaign_quality)

Call:
lm(formula = collected_funds ~ campaign_quality + social + pitch_size +
creators, data = df)

Residuals:
Min 1Q Median 3Q Max
-5913 -1508 -824 467 51945

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 719.5823 163.4679 4.402 1.11e-05 ***
campaign_quality 131.1866 22.2175 5.905 3.93e-09 ***
social 653.2001 127.2642 5.133 3.04e-07 ***
pitch_size 0.7572 0.1479 5.120 3.25e-07 ***
creators 219.9207 27.4981 7.998 1.79e-15 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 2991 on 2995 degrees of freedom

Multiple R-squared: 0.05728, Adjusted R-squared: 0.05602
F-statistic: 45.5 on 4 and 2995 DF, p-value: < 2.2e-16
summary(model_engagement)

Call:
lm(formula = comments_count ~ campaign_quality + social + pitch_size +
creators, data = df)

Residuals:

BE603: Data Analytics III 5 Version: Fall 2023

 Min 1Q Median 3Q Max
-32.99 -10.22 -4.94 3.46 724.92

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 4.573736 1.318590 3.469 0.00053 ***
campaign_quality 0.825710 0.179214 4.607 4.25e-06 ***
social 4.340921 1.026558 4.229 2.42e-05 ***
pitch_size 0.003475 0.001193 2.913 0.00361 **
creators 1.441106 0.221809 6.497 9.56e-11 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 24.13 on 2995 degrees of freedom

Multiple R-squared: 0.03344, Adjusted R-squared: 0.03215
F-statistic: 25.91 on 4 and 2995 DF, p-value: < 2.2e-16
summary(model_mediation)

Call:
lm(formula = collected_funds ~ campaign_quality + comments_count +
social + pitch_size + creators, data = df)

Residuals:
Min 1Q Median 3Q Max
-26106.6 -1106.2 -523.6 430.7 30670.5

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 362.4976 127.2558 2.849 0.00442 **
campaign_quality 66.7211 17.3222 3.852 0.00012 ***
comments_count 78.0729 1.7599 44.361 < 2e-16 ***
social 314.2919 99.1684 3.169 0.00154 **
pitch_size 0.4859 0.1151 4.222 2.49e-05 ***
creators 107.4095 21.5138 4.993 6.30e-07 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 2324 on 2994 degrees of freedom

Multiple R-squared: 0.4312, Adjusted R-squared: 0.4302
F-statistic: 453.9 on 5 and 2994 DF, p-value: < 2.2e-16
#| results: "hide"
#| warning: false
#| message: false

# You may want to display all the results side by side

stargazer::stargazer(
# The models we want to display
model_campaign_quality, model_engagement, model_mediation,

# html file, save in top_table

type = 'html',
out = "results_side_by_side.html")

<table style="text-align:center"><tr><td colspan="4" style="border-bottom: 1px sol

id black"></td></tr><tr><td style="text-align:left"></td><td colspan="3"><em>Depen
dent variable:</em></td></tr>

BE603: Data Analytics III 6 Version: Fall 2023

<tr><td></td><td colspan="3" style="border-bottom: 1px solid black"></td></tr>
<tr><td style="text-align:left"></td><td>collected_funds</td><td>comments_count</t
d><td>collected_funds</td></tr>
<tr><td style="text-align:left"></td><td>(1)</td><td>(2)</td><td>(3)</td></tr>
<tr><td colspan="4" style="border-bottom: 1px solid black"></td></tr><tr><td style
="text-align:left">campaign_quality</td><td>131.187^***</td><td>0.826<su
p>***</sup></td><td>66.721^***</td></tr>
<tr><td style="text-align:left"></td><td>(22.218)</td><td>(0.179)</td><td>(17.322)
</td></tr>
<tr><td style="text-align:left"></td><td></td><td></td><td></td></tr>
<tr><td style="text-align:left">comments_count</td><td></td><td></td><td>78.073<su
p>***</sup></td></tr>
<tr><td style="text-align:left"></td><td></td><td></td><td>(1.760)</td></tr>
<tr><td style="text-align:left"></td><td></td><td></td><td></td></tr>
<tr><td style="text-align:left">social</td><td>653.200^***</td><td>4.341
^***</td><td>314.292^***</td></tr>
<tr><td style="text-align:left"></td><td>(127.264)</td><td>(1.027)</td><td>(99.168
)</td></tr>
<tr><td style="text-align:left"></td><td></td><td></td><td></td></tr>
<tr><td style="text-align:left">pitch_size</td><td>0.757^***</td><td>0.0
03^***</td><td>0.486^***</td></tr>
<tr><td style="text-align:left"></td><td>(0.148)</td><td>(0.001)</td><td>(0.115)</
td></tr>
<tr><td style="text-align:left"></td><td></td><td></td><td></td></tr>
<tr><td style="text-align:left">creators</td><td>219.921^***</td><td>1.4
41^***</td><td>107.409^***</td></tr>
<tr><td style="text-align:left"></td><td>(27.498)</td><td>(0.222)</td><td>(21.514)
</td></tr>
<tr><td style="text-align:left"></td><td></td><td></td><td></td></tr>
<tr><td style="text-align:left">Constant</td><td>719.582^***</td><td>4.5
74^***</td><td>362.498^***</td></tr>
<tr><td style="text-align:left"></td><td>(163.468)</td><td>(1.319)</td><td>(127.25
6)</td></tr>
<tr><td style="text-align:left"></td><td></td><td></td><td></td></tr>
<tr><td colspan="4" style="border-bottom: 1px solid black"></td></tr><tr><td style
="text-align:left">Observations</td><td>3,000</td><td>3,000</td><td>3,000</td></tr
>
<tr><td style="text-align:left">R²</td><td>0.057</td><td>0.033</td><td>
0.431</td></tr>
<tr><td style="text-align:left">Adjusted R²</td><td>0.056</td><td>0.032
</td><td>0.430</td></tr>
<tr><td style="text-align:left">Residual Std. Error</td><td>2,991.197 (df = 2995)<
/td><td>24.128 (df = 2995)</td><td>2,323.911 (df = 2994)</td></tr>
<tr><td style="text-align:left">F Statistic</td><td>45.495^*** (df = 4;
2995)</td><td>25.907^*** (df = 4; 2995)</td><td>453.879^*** (d
f = 5; 2994)</td></tr>
<tr><td colspan="4" style="border-bottom: 1px solid black"></td></tr><tr><td style
="text-align:left"><em>Note:</em></td><td colspan="3" style="text-align:right"><su
p>*</sup>p<0.1; ^**p<0.05; ^***p<0.01</td></tr>
</table>
# Open the project folder: you will find the results in results_side_by_side.html
model_engagement <- lm(comments_count ~ campaign_quality,
data = df)

model_mediation <- lm(collected_funds ~ campaign_quality + comments_count,

data = df)

BE603: Data Analytics III 7 Version: Fall 2023

model_bootstrap <- mediation::mediate(model.m = model_engagement,
model.y = model_mediation,
treat = "campaign_quality",
mediator = "comments_count",
boot = TRUE,
sims = 500)
Running nonparametric bootstrap
summary(model_bootstrap)

Causal Mediation Analysis

Nonparametric Bootstrap Confidence Intervals with the Percentile Method

Estimate 95% CI Lower 95% CI Upper p-value

ACME 72.646 37.314 104.75 <2e-16 ***
ADE 78.978 33.063 125.11 0.004 **
Total Effect 151.624 106.471 192.93 <2e-16 ***
Prop. Mediated 0.479 0.264 0.73 <2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Sample Size Used: 3000

Simulations: 500
plot(model_bootstrap)

BE603: Data Analytics III 8 Version: Fall 2023

Looking at the table, as we have small p-values (<0,01) throughout we may conclude the following:
• Step 1: Campaign quality (X) has a significant total effect on collected funds (Y). For every
unit increase in campaign quality the collected funds increase on average by 151.6 dollars.
• Step 2: Campaign quality (X) has a significant effect on comments count (M). For every
unit increase in campaign quality the level of engagement increase on average by 0.9074
• Step 3: The comments count (M) has a significant effect on collected funds (Y), where one
unit increase in comments (all else equal) increase collected funds on average by 80.062
dollars. This indicates that we have a partial mediation. Also the campaign quality (X) has a
significant direct effect on collected funds (Y). For every unit increase in campaign quality
(all else equal) the collected funds increase on average by 78.978 dollars. This tells us that
there is a partial mediation - as there’s significant direct effect of the independent variable
on the dependent variable alongside a significant indirect effect through the mediator
variable and the effect size between independent and dependent variable is lower when
mediating variable is not in the model.
• Step 4: The Average Causal Mediation Effect (ACME) at 72.646 signifies the impact
channeled through engagement levels, illustrating a substantial mediated effect on collected
funds. This effect is statistically significant, underlining the influential role of engagement
in the relationship between campaign quality and funds collected. Simultaneously, the

BE603: Data Analytics III 9 Version: Fall 2023

 Average Direct Effect (ADE) of 78.978 portrays the direct influence of campaign quality on
funds collected. The Total Effect, encompassing both direct and mediated impacts, is
estimated at 151.624, further confirming the robust relationship between campaign
quality, engagement levels, and the funds collected. All the effects are significant with p-
value < 0.01.

6.1.1 Adjusting the regression model with control variables

• As evident in our table, we have a low adjusted R-squared for our mediated model (0,42).
This means that the explanatory power of our model is weak although it is significant. This
further suggests that we may be able to improve the fit by including control variables.
However, if the model does not increase in explanatory power when these additional
variables are included, we can strengthen the robustness of our conclusion.
model_campaign_quality1 <- lm(collected_funds ~ campaign_quality + social + pitch_
size + creators,
data = df)

model_engagement1 <- lm(comments_count ~ campaign_quality + social + pitch_size +

creators,
data = df)

model_mediation1 <- lm(collected_funds ~ campaign_quality + comments_count + socia

l + pitch_size + creators,
data = df)
# You may want to display all the results side by side
stargazer::stargazer(
# The models we want to display
model_campaign_quality1, model_engagement1, model_mediation1,

# html file, save in top_table

type = 'html',
out = "results_side_by_side_control.html")

# Open the project folder: you will find the results in results_side_by_side.html

As is evident by a comparison of our two tables, the pitch size and social control variables have a
significant relationship with collected funds. However, the explanatory power of the mediation
model only increases by a slight amount from 42% to 42.5%. This suggests that they should not be
included as they may contribute to over fitting.

6.1.2 Summary of hypothesis 1

Our analysis supports Hypothesis 1, affirming a substantial relationship between crowdfunding
campaign quality and the total funds collected, mediated through the level of engagement exhibited
by campaign backers. The findings underscore the importance of both direct and indirect effects in
influencing the funds collected, emphasizing the role of engagement alongside campaign quality.
Furthermore, these results suggest a partial mediation model, highlighting the coexistence of direct
and mediated effects in explaining the relationship between campaign quality, engagement levels, and
funds collected.

BE603: Data Analytics III 10 Version: Fall 2023

7. Hypothesis 2

There is a significant relationship between crowdfunding campaign quality and the level of
engagement of the backers in the campaign.

7.0.1 Variable Selection

• Dependent variable (Y): comments_count (engagement level)
• Independent variable (X): campaign_quality (Composite variable indicating campaign
quality)
model_h2 <- lm(comments_count ~ campaign_quality + social + pitch_size + creators,
data = df)

summary(model_h2)

Call:
lm(formula = comments_count ~ campaign_quality + social + pitch_size +
creators, data = df)

Residuals:
Min 1Q Median 3Q Max
-32.99 -10.22 -4.94 3.46 724.92

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 4.573736 1.318590 3.469 0.00053 ***
campaign_quality 0.825710 0.179214 4.607 4.25e-06 ***
social 4.340921 1.026558 4.229 2.42e-05 ***
pitch_size 0.003475 0.001193 2.913 0.00361 **
creators 1.441106 0.221809 6.497 9.56e-11 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 24.13 on 2995 degrees of freedom

Multiple R-squared: 0.03344, Adjusted R-squared: 0.03215
F-statistic: 25.91 on 4 and 2995 DF, p-value: < 2.2e-16
#| results: "hide"
#| warning: false
#| message: false

# You may want to display all the results side by side

stargazer::stargazer(
# The models we want to display
model_h2,

# html file, save in top_table

type = 'html',
out = "results_side_by_side.html")

<table style="text-align:center"><tr><td colspan="2" style="border-bottom: 1px sol

id black"></td></tr><tr><td style="text-align:left"></td><td><em>Dependent variabl
e:</em></td></tr>
<tr><td></td><td colspan="1" style="border-bottom: 1px solid black"></td></tr>

BE603: Data Analytics III 11 Version: Fall 2023

<tr><td style="text-align:left"></td><td>comments_count</td></tr>
<tr><td colspan="2" style="border-bottom: 1px solid black"></td></tr><tr><td style
="text-align:left">campaign_quality</td><td>0.826^***</td></tr>
<tr><td style="text-align:left"></td><td>(0.179)</td></tr>
<tr><td style="text-align:left"></td><td></td></tr>
<tr><td style="text-align:left">social</td><td>4.341^***</td></tr>
<tr><td style="text-align:left"></td><td>(1.027)</td></tr>
<tr><td style="text-align:left"></td><td></td></tr>
<tr><td style="text-align:left">pitch_size</td><td>0.003^***</td></tr>
<tr><td style="text-align:left"></td><td>(0.001)</td></tr>
<tr><td style="text-align:left"></td><td></td></tr>
<tr><td style="text-align:left">creators</td><td>1.441^***</td></tr>
<tr><td style="text-align:left"></td><td>(0.222)</td></tr>
<tr><td style="text-align:left"></td><td></td></tr>
<tr><td style="text-align:left">Constant</td><td>4.574^***</td></tr>
<tr><td style="text-align:left"></td><td>(1.319)</td></tr>
<tr><td style="text-align:left"></td><td></td></tr>
<tr><td colspan="2" style="border-bottom: 1px solid black"></td></tr><tr><td style
="text-align:left">Observations</td><td>3,000</td></tr>
<tr><td style="text-align:left">R²</td><td>0.033</td></tr>
<tr><td style="text-align:left">Adjusted R²</td><td>0.032</td></tr>
<tr><td style="text-align:left">Residual Std. Error</td><td>24.128 (df = 2995)</td
></tr>
<tr><td style="text-align:left">F Statistic</td><td>25.907^*** (df = 4;
2995)</td></tr>
<tr><td colspan="2" style="border-bottom: 1px solid black"></td></tr><tr><td style
="text-align:left"><em>Note:</em></td><td style="text-align:right">^*p<0
.1; ^**p<0.05; ^***p<0.01</td></tr>
</table>
# Open the project folder: you will find the results in results_side_by_side.html

The output from the model shows that there is a positive and significant relationship between the
campaign quality and comments count, p-value< 0.05, thus we reject the null hypothesis. A one unit
increase in campaign quality gives a 0.99 unit increase in comments count. We support hypothesis 2
that there is a significant relationship between campaign quality and the engagement level of the
backers such that a one unit increase in quality increases the engagement of the backers by 0,99 units.
The explanatory power of the model is low, given an adjusted R-squared of 0.019.

8. Hypothesis 3

The relationship between crowdfunding campaign quality and the level of engagement of the backers
in the campaign is moderated by the amount of effort the campaigners engage to interact with the
backers.

8.0.1 Variable Selection

• Dependent variable (Y): comments_count (engagement level)
• Independent variable (X): campaign_quality (Composite variable indicating campaign
quality)
• Moderator (M): updates_count (efforts the campaigners engage to interact with the
backers)

BE603: Data Analytics III 12 Version: Fall 2023

model_h3 <- lm(comments_count ~ campaign_quality * updates_count + social + pitch_
size, data = df)

summary(model_h3)

Call:
lm(formula = comments_count ~ campaign_quality * updates_count +
social + pitch_size, data = df)

Residuals:
Min 1Q Median 3Q Max
-87.81 -9.44 -4.46 3.74 709.23

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 4.890297 1.334751 3.664 0.000253 ***
campaign_quality 0.933453 0.188972 4.940 8.26e-07 ***
updates_count 2.476165 0.259440 9.544 < 2e-16 ***
social 3.606693 1.004682 3.590 0.000336 ***
pitch_size 0.002021 0.001173 1.722 0.085138 .
campaign_quality:updates_count -0.224840 0.031805 -7.069 1.93e-12 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 23.59 on 2994 degrees of freedom

Multiple R-squared: 0.07627, Adjusted R-squared: 0.07473
F-statistic: 49.44 on 5 and 2994 DF, p-value: < 2.2e-16
#| results: "hide"
#| warning: false
#| message: false

# You may want to display all the results side by side

stargazer::stargazer(
# The models we want to display
model_h3,

# html file, save in top_table

type = 'html',
out = "results_side_by_side.html")

<table style="text-align:center"><tr><td colspan="2" style="border-bottom: 1px sol

id black"></td></tr><tr><td style="text-align:left"></td><td><em>Dependent variabl
e:</em></td></tr>
<tr><td></td><td colspan="1" style="border-bottom: 1px solid black"></td></tr>
<tr><td style="text-align:left"></td><td>comments_count</td></tr>
<tr><td colspan="2" style="border-bottom: 1px solid black"></td></tr><tr><td style
="text-align:left">campaign_quality</td><td>0.933^***</td></tr>
<tr><td style="text-align:left"></td><td>(0.189)</td></tr>
<tr><td style="text-align:left"></td><td></td></tr>
<tr><td style="text-align:left">updates_count</td><td>2.476^***</td></tr
>
<tr><td style="text-align:left"></td><td>(0.259)</td></tr>
<tr><td style="text-align:left"></td><td></td></tr>
<tr><td style="text-align:left">social</td><td>3.607^***</td></tr>
<tr><td style="text-align:left"></td><td>(1.005)</td></tr>
<tr><td style="text-align:left"></td><td></td></tr>

BE603: Data Analytics III 13 Version: Fall 2023

<tr><td style="text-align:left">pitch_size</td><td>0.002^*</td></tr>
<tr><td style="text-align:left"></td><td>(0.001)</td></tr>
<tr><td style="text-align:left"></td><td></td></tr>
<tr><td style="text-align:left">campaign_quality:updates_count</td><td>-0.225^***</td></tr>
<tr><td style="text-align:left"></td><td>(0.032)</td></tr>
<tr><td style="text-align:left"></td><td></td></tr>
<tr><td style="text-align:left">Constant</td><td>4.890^***</td></tr>
<tr><td style="text-align:left"></td><td>(1.335)</td></tr>
<tr><td style="text-align:left"></td><td></td></tr>
<tr><td colspan="2" style="border-bottom: 1px solid black"></td></tr><tr><td style
="text-align:left">Observations</td><td>3,000</td></tr>
<tr><td style="text-align:left">R²</td><td>0.076</td></tr>
<tr><td style="text-align:left">Adjusted R²</td><td>0.075</td></tr>
<tr><td style="text-align:left">Residual Std. Error</td><td>23.591 (df = 2994)</td
></tr>
<tr><td style="text-align:left">F Statistic</td><td>49.444^*** (df = 5;
2994)</td></tr>
<tr><td colspan="2" style="border-bottom: 1px solid black"></td></tr><tr><td style
="text-align:left"><em>Note:</em></td><td style="text-align:right">^*p<0
.1; ^**p<0.05; ^***p<0.01</td></tr>
</table>
# Open the project folder: you will find the results in results_side_by_side.html

We did not standardize the variables we used because they are discrete, not continuous variables,
comments_count, campaign_quality, updates_count. It is for example not possible for comments
count to take on a value of 1.2.
The output from the model shows a negative significant moderating effect between updates count
and comments count, p-value< 0.05, thus we reject the null hypothesis. We support hypothesis 3
that the relationship between crowdfunding campaign quality and the level of engagement of the
backers in the campaign is moderated by the amount of effort the campaigners engage to interact
with the backers is significant. The explanatory power of the model is low, given an adjusted R-
squared of 0.075.

9. Summary

We reject the null for all three hypothesis, thus our models shows statistical significant support for
the three hypothesis we tested. However, none of the models had a large explanatory power, using
0.6 as a benchmark, none of the models had an adjusted R-squared that exceeded this benchmark.

BE603: Data Analytics III 14 Version: Fall 2023