- Mayanka

# Free Hugs T-shirt Will Get You More Free Hugs than the Free Hugs Signboard

Updated: Dec 18, 2017

**Introduction **

Have you ever wondered if you will get free hugs from strangers? There have been “Free Hugs” campaigns around the world as a social movement involving individuals who offer hugs to strangers in public places. The hugs are meant to be random acts of kindness—selfless acts performed just to make others feel better.

The campaign in its present form was started in 2004 by an Australian man known only by the pseudonym "Juan Mann". The campaign became famous internationally in 2006 as the result of a music video on YouTube by the Australian band Sick Puppies, which has been viewed over 77 million times as of November 20, 2015. The first International Free Hugs Day was July 7, 2007, the second on July 5, 2008, and the third was on July 4, 2009. Since then International Free Hugs Month is celebrated on the first Saturday of July and continues until August first. Continuing the nonviolent movement of Dr. Martin Luther King, Jr., the mission of the Free Hugs Project is to spread love, inspire change and raise awareness of social issues.

But the question is which works better for the Free Hugs Campaign, T-shirt or Signboard? For quite some time, the myth involving around this social campaign is, “T-shirt wearer tends to get more free hugs than the sign holder”. In this project, we have studied and analyzed this myth.

**Study Method**

Our method of study is experimental with convenience sampling. We have observed the number of hugs received by the Myth-busters while wearing Free Hugs T-shirt versus holding the Free Hugs signboard at Stamp Union, University of Maryland campus.

**Population:** Our population is everyone who hangs out at the Stamp Union on the campus. **Sample:** Our sample is the people who have hugged the myth-busters on the day during that specific time when the myth-busters were collecting their data.

Hypothesis: Wearing a T-shirt with “Free Hugs” quote written on it will get you more free hugs from public in comparison to the hugs received by “Free Hugs” sign holder. Operationalized Hypothesis: If you stand at the University of Maryland campus wearing a Free Hugs T-shirt, you will get more number of hugs when compared to standing with the free hugs signboard.

**Data Collection **

The data was collected at Stamps Student Union on Tuesday, Nov 7 between 11:30am – 1:00pm. The myth-busters spent 60 minutes at the location in each mode of Free Hugs, i.e. 60 mins wearing the T-shirt and 60 mins with the signboard. There was a 20 mins break between the modes of free hugs in order to maintain the independence of the sample.

Total Number of Observation = 24, 12 for each mode of free hugs

**Independent variable:** “Medium of Free Hugs”, i.e. wearing Free Hugs T-shirt or using Free Hugs signboard - nominal scale with two levels, thus we can use pie chart and bar plot to show the proportion.

**Dependent Variable:** Number of Hugs received in each mode – ratio scale

**Parameter:** The mean number of hugs myth-busters will receive at Stamp Union for each mode of free hugs is our parameter.

**Statistics:** The mean number of hugs received by our myth-busters for our sample is the statistics.

**Analysis **

We are trying to analyze if there is a significant difference between the population means of free hugs received if the Myth-buster wears a Free Hug T-shirt vs holding a sign-board. Since, our independent variable, “Medium of Free Hugs” is a nominal variable with two levels and our Dependent Variable, “Number of Hugs received” during different time phases is a ratio scale, the Two Sample T-Test best fits our study question.

Null Hypothesis, Ho = The mean population of number of hugs received by myth-busters while wearing a Free Hugs t-shirt will be equal to the mean population of number of hugs received while standing with a free hugs signboard, i.e., 𝜇(𝑡𝑠ℎ𝑖𝑟𝑡) = 𝜇(𝑠𝑖𝑔𝑛𝑏𝑜𝑎𝑟𝑑) Alternate Hypothesis, Ha = The mean population of number of hugs received by myth-busters while wearing a Free Hugs t-shirt will not be equal to the mean population of number of hugs received while standing with a free hugs signboard, i.e., 𝜇(𝑡𝑠ℎ𝑖𝑟𝑡) ≠ 𝜇(𝑠𝑖𝑔𝑛𝑏𝑜𝑎𝑟𝑑)

Since the p-value= 0.14>α (0.05), we fail to reject the null hypothesis. This is the probability of obtaining a sample statistic as extreme as ours under the null hypothesis and we do not have enough evidence to reject the null hypothesis. Hence, we observe that there is no statistically significant difference between the mean population of number of hugs received during the two categories of free hugs collection.

As there is no statistically significant difference, we won’t be calculating effect size. However, we will calculate the power at three levels of Cohen’s d of 0.2, 0.4 and 0.8 which represents small, medium and large effect sizes respectively for our test. From the table below, we can say that given our sample size we would have had 10% chance of detecting a small effect, 27% chance of detecting a medium effect, and 77% chance of detecting a large effect.

**Conclusion ** Based on our study, we observe that there is no statistically significant difference between the mean population of number of hugs received during the two categories of Medium of free hugs. (t = -1.56 and p =0.14). Since we do not have enough evidence to reject null hypothesis, we can’t conclude anything about our Myth. There might be other factors affecting our study and statistical test. Hence, the study needs to be done again. In future, we should understand the study design properly and try to collect sample data efficiently including more factors such as Age and time of the day. We recommend collecting the sample data on different dates in the university campus.

**Test of Assumptions**

Like any other statistical tests, two sample t-test has certain assumptions. We conduct test of assumptions to make sure that the sample data does not violate these assumptions and gives us a valid result.

**Homogeneity of variance**- On performing Levene’s test, we get p-value > 0.05. Hence, we observe no violation of homogeneity of variance.**Normality**– Our dependent variable should be approximately normally distributed for each group of the independent variable. From the Histogram and QQ-plot graph above, we observe a violation of normality.**Independence of Observation**– There should be no relationship between the observations in each group or between the groups themselves. Since our data collection method was not robust and there might be a violation of this assumption.

**Limitations **

Our sample size was very small even after dividing the number of hugs in 10 minutes interval.

Normally, we should have replicated this experiment at least 12 times or more to see stronger relationship.

The assumption of Independence of Observation is being violated.

The samples should be collected using simple random sampling. However, our data has convenience bias which makes our sample non-representative of the population.

As it was raining when we collected our data, the specific weather condition might have influenced people’s willingness and mood for hugs.

Also, the familiarity that the experiment was for a course might have affected fellow students to hug the myth-busters.

External factors which possibly could have affected the mood of the public prior to sample collection for instance exam day, quiz day or family problems.

**Lessons Learnt **

In future, we should understand the study design prior to data collection so that we can randomize our sample and avoid experimental errors.

We should have repeated the data collection process to collect larger sample. Ideally, we should have collected our sample data on at least 3-4 different days.

We also need to collect data efficiently including more factors such as age, time of the day, etc.

We should have collected sample data from different location on campus as well. As many people might not go to Stamp Union.

We should include more myth-busters

A significant amount of time interval should be allowed between the two modes of free hugs while collecting the sample. This will ensure independence of observations.