Introduction
Our class group conducted a study to analyze smartphone use and social use and social media habits of people we could easily contact. The sample was therefore a convenient sample. The people surveyed were all from the population of American adults. The survey questions were conducted through phone, email, text message, social media, and face to face interviews.
Responses were voluntary, hence not everyone approached actually responded to the survey questions.
The following questions were asked:
1. How many hours a day do you spend on the smart phone (estimate to the nearest hour)?
2. How many times a day do you visit or use social media (Facebook, Instagram, Twitter, etc)?
3. Do you use your smart phone for work? Yes/No_______
4. Which social media site do you use most often? Facebook/Instagram/Twitter/Other_____
The results from 121 people who responded are summarized in the table below:
Looking at the Categorical Variables
The responses to the question "Do you use your smart phone for work?" are illustrated in the pie chart below.
The pie chart shows that more people, a slight majority of the respondents (about 55%), use smartphones for work. About 44% of respondents do not use smartphone for work.
To look at the relationship between smartphone use at work and frequency and preference for social media sites, the bar plot below is used:
The bar plot above illustrates that Facebook was the most frequently visited social media site by both groups who used smartphone for work and those who did not. Comparatively, Facebook was more frequently accessed by those who did not use smartphone for work (approximately 41%) against those who used smartphone at work (approximately 31%).
For those who used smartphone at work, only 4 out of 121 respondents frequently visited Twitter, the least used site in this particular lot. Among those who did not use smartphone at work, the least frequented social media site was Instagram, by about 4 out of 121 respondents.
Some participants reported using "other" social media sites. While collecting data, some respondents who reported no frequent use of social media were also included in this category during tabulation of results. This calls for careful drawing of conclusions and interpretation of the data.
Our study would have been better if we included a separate category of "none". What is evident is that there was a relatively small percentage of people who either did not use social media at all, or used sites other than Facebook, Instagram and Twitter among both groups of smartphone users at work (5%) and those who did not (5.8%). A big majority in both groups of smartphone users at work (95%) and nonusers (94.2%) frequented at least one social media site on a daily basis.
Social media usage was therefore prevalent irrespective of whether one used smartphone at work or not. There was also not much significant difference in social media preference between those who use smartphone at work and those who do not  Facebook being the more popular site among both groups.
Looking at a Numerical Variable
The responses to the question "How many hours a day do you spend on the smartphone (estimate to the nearest hour)?" are depicted in the histogram, box plot and summary statistics below:
Summary statistics:

The histogram shows a unimodal and bellshaped distribution that is skewed to the right. The modal class is 2 to 3 hours. The spread, which is also represented by the range, is 7 hours. The histogram distribution reveals a gap in the modal class between 6 and 7 hours. This may suggest that the value to the right of the gap, 8 hours per day of smartphone use, is an outlier.
The box plot clearly illustrates that the 8 hours/day value is indeed an outlier, depicted as a lone dot outside the limit of usual distribution. The box plot also shows that the median is 2 hours/day of smartphone use, which is also shown in the summary statistics. Comparatively, the mean is 2.57 hours/day of smartphone use.
The center of a skewed data set is best described by the median, which is not affected by extreme values or outliers. In this case, the mean being bigger than the median is because of the rightskewed nature of the distribution. The extreme outlier tends to push the mean further to the right, while the median is resistant to outliers.
The midrange, (1 + 8)/2, is 4.5 hours/day, which is way off compared to the mean (2.57 hours/day) or median (2 hours/day), and thus is not a reasonable estimate of the center.
The standard deviation is approximately 1.4 hours which indicates that there is not a huge variability in the responses to this question.
The Inter Quartile Range (IQR), the range of the middle half of the data, is 2 hours. This is comparable to the median.
The range rule of thumb (range/4) is = 7/4 = 1.75 hours/day. This is a fair estimate of the standard deviation (1.4 hours) for this particular data. The slight overestimation of the standard deviation can be explained by the rightskewedness of the distribution and the presence of an outlier at 8 hours.
The outlier of 8 hours/day spent on smartphone does not appear to be an error. It is quite likely a legitimate data value because it is possible for someone to spend 8 hours a day on smartphone, especially if using it for work.
Overall, these findings show plausible use of smartphones, with typical use of 2 to 3 hours a day (modal class) which does not appear to be erroneous.
Looking for a Relationship Between Two Numerical Variables
To examine if there is a relationship between the responses to the questions "How many hours a day do you spend on the smartphone?" and "How many times a day do you visit/use social media?" the scatter plot of the paired data was analyzed.
The scatter plot shows that there is a weak positive and linear correlation between hours/day spent on smartphone and the frequency/day of visiting social media. It is weak because of the widespread scatter of the dots, but it is still discernible to see an upward positive trend from the scattered dots.
A clear outlier can be seen at (8 hours/day, 22 visits/day). Such an outlier may significantly affect the correlation coefficient. StatCrunch technology reveals that the correlation coefficient for the paired data is 0.526 (rounded to the nearest thousandths):
Correlation between Hours/day and Frequency/day is: 0.52565759 
Using an estimated sample size of n=100 (actual sample size is 121), the critical value from Table A5 of the textbook is 0.196. When compared to the absolute value of the computed correlation coefficient from StatCrunch, 0.526, the latter is bigger than 0.196.
It can therefore be concluded that there is sufficient statistical evidence to support the claim of linear correlation between the number of hours a day spent on the smartphone and the number of times a day social media is visited. However, that correlation may not be a strong one as suggested from the scatter plot and the 0.526 correlation coefficient. A correlation coefficient: near zero (0) suggests no correlation, near one (1) suggests strong correlation, while in the middle near 0.5 suggests weak correlation.
Already a member? Sign in.