Landkit Redesign

We design & build products, apps, and sites for companies doing good things for the planet.

Landkit Redesign

We design & build products, apps, and sites for companies doing good things for the planet.

Social Media and Donation

A quantitative study to understand donor willingness on social media

In the last decade, social media has become an available tool for online charity donations. Facebook, Twitter, GoFundMe.com are few of the platforms on which individuals or communities can partake in fundraising efforts. Our research aims to further expand how the social network effect, an organization’s network of followers, play a huge role in the success of charitable donations where the circle of online friends in social networks helps promote causes (Saxton and Wang, 2014). This study expands existing research by identifying how an individual’s social circle affect their willingness to donate. This research may drive how non-profits or other charities can navigate and target social media strategies to increase charity donations.

How might seeing charity donations from different users on social newsfeed impact a person’s willingness to donate to charity?

Variables

IV : Users (Friends, Influencers, Strangers (Control)

We included two variables to serve as influencers in an individual's social network. The first variable was Facebook friends. Friends from the participant’s Facebook profile were included on the prototype and listed on the feed through various activities like donating to a cause, going to a concert or an event. The second variable was famous celebrities and philanthropists. Examples include a popular actor like Leonardo Dicaprio donating for a cause. The control group only included strangers in the newsfeed where we showed a random person on the feed who donated recently for a cause.

DV : Donation Willingness, Donation Amount

Our dependent variable, donation willingness was measured by the count of button clicks and donation amount in dollars among each participant in the three groups: famous celebrities, friends, and strangers.

Hypotheses

  • H1 (Hypothesis): The presence of donations from friends increases likelihood of participant’s action to donate.

  • H0 (Null Hypothesis): The presence of donations from friends will not change likelihood of participant’s action to donate.

Task Design

The experiment used the concept of using a fictional payments app called PayNoww. A high fidelity interactive Invision alternative prototype was created to simulate the experience of booking events and donating to a cause on the home feed (social newsfeed). We distributed a Google Sites page with the prototype along with pre and post survey links to each participant. Each participant’s PayNoww account was credited with USD 30 which was informed at the beginning of the experiment.

The prototype was modified for each participant group:

  • The stranger's group was given a prototype where the feed showed strangers sending money and booking event tickets and donating to a cause (Syrian war).

  • The celebrity group had an celebrity, in this case, Leonardo Dicaprio, a well-known philanthropist, donating to a cause (Syrian war).

  • The friend's group’s prototype was modified by including a friend from their Facebook friends list donating to a cause (Syrian war).

Two different surveys were included in this experiment. Each participant was required to participate in a pre-survey followed by interaction with the Invision prototype. The participant’s interactions with the prototype were tracked through an analytic tool called Maze. At the end of the session, each participant was given a code that they had to paste in the post-survey to validate if they completed the usability study of the prototype. After the completion of the usability study, participants were requested to take a post-survey.

  • The Pre-Survey collected demographic information. The participants were asked to provide their gender, age, ethnicity, level of education, marital status, and family income.

  • The Post-Survey asked participants how much amount they donated out of the USD 30 provided in their virtual wallet.

Participants’ responses were filtered and curated by following parameters:

  • Each participant account could only submit one response to one task

  • Each IP address can only submit one response to one task

  • Participants were restricted within groups to U.S. and Indian citizens over the age of 18 years

  • Participants were randomly assigned groups to test the survey and prototype

...
...
...

Participants and recruitment

We initially tried using Amazon's Mechanical Turk, however, owing to the platform's privacy restrictions. We could not ask participants with links to their social media profiles (so that we can use them to customize the prototype). As a result we recuited participants from our social circles.


...

Procedure

During the experiment, our aim was to collect 10 responses for each task. We simply contacted participants for the study providing them a link to a google site which comprised of (1) Instructions (2) the pre-survey (3) Maze prototype link and (4) the post survey. Each participant was randomly assigned the groups and we had 3 different sites for each group.


...

Using Maze to see user behavior

Maze Design is a cloud based prototype analytics tool, which tracks participants’ interactions with the prototype. It helps to define and collect actionable insights. In this experiment Maze was used to verify if the participants are truly going through the prototype. It helped us in eliminating any outliers in the study. It was also interesting to see how many clicks were performed and how much time it took for participants to go through the prototype.

... ...

Method Selection and Results

The experiment examined prototype interactions of 30 participants. These 30 participants were divided in three groups of 10. For these three groups our were testing was based on how friends, strangers and celebrity influence willingness to donate. Later, we examined how much amount participants in each group donated. We interpreted that we will require two separate tests. One which will calculate the willingness to donate among the three groups and another for analyzing the donation amount. We used Chi-square test for willingness to donate and One-Way ANOVA to analyze the donation amount.

Variables
3

Independent Variable: Donations from different users

Nominal - Discrete with Three Categories

3 levels: friends, influencers, strangers - control group

Dependent Variable: User donation willingness

Dichotomous - Discrete

2 levels: Yes or no

Dependent Variable: Donation Amount

Ratio - Continuous

Statistical Methods
2

Chi Square Test (DV: User Donation Willingness)

IVs and DVs are Nominal/Categorical

One-Way ANOVA (DV: Donation Amount)

IV is Nominal/Categorical

DV is Ratio/Continuous

...

Chi-Square test for User Donation Willingness

p-value = 0.62 (Not Significant!)

There was a higher participants count of those that donated in the influence group than those who were in control group.
Control - 3 count, Friends - 2 count, Influencer - 4 count

...

One-Way Anova for Donation Amount

P-value = 0.71 (Not Significant!)

Participants in the Friends group donated more amount.

Limitations and Discussion

This study investigated social media and willingness to donate. The descriptive survey research design was adopted in the study with the interactive mockup PayNoww. A sample of 30 respondents was selected from a population. A simple random sampling technique was used for the selection, where ten participants for the each group was selected. The data collected were analyzed with the use of descriptive statistics of frequency count and percentage, and the hypotheses were tested using the inferential statistics of Chi-square and One-Way Anova at 0.05 level of significance. The result from the findings of this study showed that, though Social media can be influential for donations, we discovered that both peer and celebrity were not found to be significant moderators on participant’s intention to donate. We would caution against interpreting our finding to mean that peer and celebrity have no moderating effect. As previous studies have point the opposite to be true (Saxton & Wang, 2014; Wymer & Drollinger, 2015) . The magnitude effect between moderating effect of peer and celebrity influence needs to be explored in future research.

In this study, we relied on Facebook friends from peers as moderating variable to inform participants of peer endorsement. We randomly selected one peer as the endorser. However, it is unknown the strength of friendship and influence of one peer stacked against multiple peers. Could two or three peer endorsers, be used in tandem for greater effectiveness?

Future research could include attitudes towards charity as an outcome variable. The prototype may have a direct effect and change on participant’s attitude towards charity. The attitude change may result in future behavioral responses to support charity such as future donation, volunteering, favorable referrals, as consequence of multiple exposures to charity’s appeals and endorsements over time. It would be then useful to measure participant responses to volunteering and referrals.

Read the full report