TheFacebook
In case you were curious, here are the results of our study:
We were looking to see wht factors on thefacebook correlated with the number of friends that a harvard undergraduate would list. We tested for year in school, major, political beliefs, number of groups they joined, gender, relationship status, hometown size, etc. we took random samples based on birthdays and ended up with a sample size of 308 active harvard undergrads (we omitted imaginary students, alumni, and grad students).
we found significant relations in concentration: those listed as math or science concentrators had significantly fewer friends than social science concentrators. Those in the arts and humanities also had siginificantly fewer friends than social science concentrators. Conputer science concentrators had significantly fewer friends than any other major that we had in our sample. Ec and Gov were statistically the same. Social Studies had the highest mean, but it wasn't significant at P<0.05.
We found a weak positive relation between ln(hometown size) and number of friends, meaning if you came from a larger city, you are more likely to have more friends listed. It was significant, but R^2 was low, meaning it accounts for very little of the variation in the number of friends.
There was (no surprise) a strong correlation between number of groups that one was in and the number of friends, but we took this as a sign for enthusiasm for thefacebook, as i'm guessing not a lot of people meet each other through groups like "the sex in widener interest group" as opposed to HRO or something. (though, who knows?)
there was no real difference between the genders. Those who listed themselves as single had fewer friends than those who listed themselves as in a relationship, but it wasn't quite significant (P<0.15). class year was also significant, as people tended to get more friends each year. interestingly, 2006 had the most friends over all.
what i thought was most interesting about our data was dealing with political affiliation. we first grouped the data into three sections by polititical beliefs: 1. very conservative and conservative 2. moderate 3. liberal and very liberal. There was very little difference between the groups. But! then we grouped the data by extremeness of political opinion, so: 1. very conservative, very liberal 2. liberal, conservative 3. moderate and found that moderates had significantly more friends than any one with political leanings, and that those with extreme views had significantly fewer friends than those with smaller leanings.
We came up with a lot of interesting interpretations for the data. i won't bore you with our conjectures. There was obviously a response bias, as not everyone fills out every field. Also, since we sampled by birthday, we clearly left out the people who didn't list their birthdays.
yeah. i definitely spend too much time on thefacebook.
We were looking to see wht factors on thefacebook correlated with the number of friends that a harvard undergraduate would list. We tested for year in school, major, political beliefs, number of groups they joined, gender, relationship status, hometown size, etc. we took random samples based on birthdays and ended up with a sample size of 308 active harvard undergrads (we omitted imaginary students, alumni, and grad students).
we found significant relations in concentration: those listed as math or science concentrators had significantly fewer friends than social science concentrators. Those in the arts and humanities also had siginificantly fewer friends than social science concentrators. Conputer science concentrators had significantly fewer friends than any other major that we had in our sample. Ec and Gov were statistically the same. Social Studies had the highest mean, but it wasn't significant at P<0.05.
We found a weak positive relation between ln(hometown size) and number of friends, meaning if you came from a larger city, you are more likely to have more friends listed. It was significant, but R^2 was low, meaning it accounts for very little of the variation in the number of friends.
There was (no surprise) a strong correlation between number of groups that one was in and the number of friends, but we took this as a sign for enthusiasm for thefacebook, as i'm guessing not a lot of people meet each other through groups like "the sex in widener interest group" as opposed to HRO or something. (though, who knows?)
there was no real difference between the genders. Those who listed themselves as single had fewer friends than those who listed themselves as in a relationship, but it wasn't quite significant (P<0.15). class year was also significant, as people tended to get more friends each year. interestingly, 2006 had the most friends over all.
what i thought was most interesting about our data was dealing with political affiliation. we first grouped the data into three sections by polititical beliefs: 1. very conservative and conservative 2. moderate 3. liberal and very liberal. There was very little difference between the groups. But! then we grouped the data by extremeness of political opinion, so: 1. very conservative, very liberal 2. liberal, conservative 3. moderate and found that moderates had significantly more friends than any one with political leanings, and that those with extreme views had significantly fewer friends than those with smaller leanings.
We came up with a lot of interesting interpretations for the data. i won't bore you with our conjectures. There was obviously a response bias, as not everyone fills out every field. Also, since we sampled by birthday, we clearly left out the people who didn't list their birthdays.
yeah. i definitely spend too much time on thefacebook.