Integrative Science Symposium: e-Relationships


We’ve done experiments and we’ve done cross sectional and longitudinal study to investigate the effects
or the relationship between social media use and self concept clarity and self esteem, social competence, quality of friendships, friendship initiation, online sexual risk-taking, cyber
bullying and many more, and you know the object object of our study is a moving target. Often times
the tools we use in our research already get outdated before our results are published
and here I give you the latest figures on social media use and as you will see social
media use has globalized enormously. Ten years ago in the Netherlands we had our own Dutch
apps and social media tools but now you see that the tools used
are international. Whatsapp as a text messaging app is by far the most popular in the Netherlands
and I think in many other countries as well. Traditional SMS is still used but only when
adolescents don’t have Wi-Fi available. And Snapchat is now the third one but
it has risen enormously in the past months. I think it’s above SMS right
now and then you see Telegram and Klick. These are social media that are used for personal
communication mostly one to one, communication and these are the social media that are used
for group communication, and you see that YouTube is the most popular one, closely followed by
Facebook, Twitter, Instagram — all international names — Google+, and also I’m sure that Instagram
has become very popular in the past months as I’ll show I think that Instagram has now
surpassed Twitter. If you want to investigate the social effects of social media or media
and technology you can basically rely on two types of theories. First
mass communication theories. And these theories arose in the 1920’s in response to the new
opportunities to reach large audiences via the mass media — press, radio, and the movies, and these theories find their origin in sociology and political communication but
you can also rely on or contribute to computer mediated communication theories, CMC theories,
and these theories arose in the 1970’s long before the internet became popular, and they
find their origins in interpersonal communication and social psychology. Both theories share
some parallels. First, both mass media and CMC theories acknowledge external factors can
influence or moderate the communication process. In the early theories these external factors
were called or named noise, but we now see these external factors as highly relevant
and they also both acknowledge that recipients of communication have some degree of agency, that is they are autonomous in how they perceive or decode media and communication. However,
both mass media and effect theories and research computer mediated communication research (CMC research) are still too often based in a reception model — the notion that certain
content or other properties of media or technologies have a unidirectional impact
on recipients. Mass media effects research is for instance still based on the idea
that certain properties of mass media such as modality, auditory, visual, or you name it,
or structure, special effects, visual surprises, or content influences, how people are effected
by media, and the same holds for CMC. theories especially The older ones still focus on the effects
of certain CMC characteristics. For example, asynchronicity — whether you can press the button
immediately or delayed, or anonymity, visual anonymity, and they think that this characteristic
effect a recipient. Basically, both types of theories have paid lots of
attention to interpersonal processes but far less to intrapersonal processes of both the
recipient and the sender, and they still do not recognize enough that in the newer media
environment, recipients of communication are are also senders of communication. That’s a phenomenon
for which the now somewhat obsolete term pro-humors was coined and what has been overlooked largely
until recently is that the senders can not only influence recipients by their communication
but also themselves. And this phenomenon, that our own online communicative behavior, influence
our self has been referred to as an online expression-effect. And in the past decades
we have often found that adolescents change not as much as a recipient of communication
but as a result of their own online communicative behavior and such online expression-effects
are of course highly relevant in adolescence. After all intrapersonal changes are normative
in adolescence. Adolescence is marked by three developmental goals. Adolescents need to form
an identity, they need to form intimacy for close friendships and romantic relationships
and they have to form a sexual self. These are are intrapersonal changes and what we have found
in the past ten years is that some of the affordances of social media, the opportunities
they offer the user, facilitate expression-effects related to the three developmental goals: identity, intimacy and sexuality, and one such affordance is reduced cues. Users of social media can
choose whether they want to have all your visual cues or not, so they can if they want
hide their pimples or a bad hair day and another affordance is edit ability — they can endlessly
optimize their visual and verbal self, presentation and asynchronicity — they can
also decide, “Do I press the send button now or do I wait awhile so I can still work on
it?” And what we found is that these affordances of social media is the importance
that adolescences attach to these affordances predicts their perceived control-ability. That’s
not control-ability objectively but their perceived control-ability. It may be an illusion
of control-ability. And control-ability means that they have the freedom to decide what,
how and to whom they communicate and this perceived control-ability in turn is related
to the three developmental goals in adolescence. Now you think, “Why is this the case? Why
is perceived control-ability related to the three developmental goals?” And that is because
perceived control-ability of online communication facilitates the practice, the rehearsal of
three developmental skills that are necessary to accomplish the developmental goals. And
now you think such a complicated process so simple in one model. Now I am going to
give you some examples of our research. In the past ten years on the three developmental
skills and I start with self-presentation. We have seen mere writing a blog for example
can change the writers self-esteem and their identity. They think that others understand
them better because of writing the blogs and that’s a direct expression effect. But expression
effect can also occur indirectly and we have seen that in teenagers’ self-presentation on
social networking sites . These social networking sites are designed so that users
give themselves positive feedback and that’s what we saw. 93% of ten to nineteen years old
mostly receive positive feedback. And that’s logical because the sites stimulate
positive feedback. What we also found is that the more
they used it the more feedback they received, which is logical too because the more they
use it the higher the chance they get feedback. But that’s not the only thing. The more feedback
they received the more they started to adjust their profile based on this feedback, and they
made it more attractive. They optimized their profile and this optimization of this profile
led to even more positive feedback and eventually it enhanced their self-esteem. So by creating
their own positive feedback they enhanced their own self-esteem. Another
example is a second developmental skill self-disclosure and that’s the skill to adequately disclose
yourself intimately to friends. And it’s a critical skill in adolescence. It has been related
to friendship formation, friendship quality, emotional support, well being etc., and in our
project we created skills to measure online and offline self disclosure to close friends
with items such as “how much do you tell about your worries,” about being in love, about “the
things you are ashamed of,” “your secrets”, etc. Here you see a graph where
we’ve plotted online and offline self-disclosure, on the left for girls and on the right
for boys. And what you see for girls is that the highest curve is for
offline self-disclosure and is lower for online but you see for girls that when they are ten
there’s a sharp increase in both online and offline intimate self-disclosure. When you compare
that with boys you see that this process starts around two years later at around thirteen
and fourteen years. But what’s more interesting although you see that for both groups offline
self-disclosure is higher than online self-disclosure, there are huge individual differences. For
example in the boys group thirty percent of boys disclose more online than offline about
intimate topics such as being in love and things they’re ashamed of. So for boys it is stronger
than it is for girls. But still twenty-two percent of girls prefer
online self-disclosure about intimate topics to offline self disclosure. Now next what we
wanted to investigate was how type of self disclosure was related, and we had two hypotheses.
One is the displacement hypothesis where thought its possible if they disclose online they
don’t disclose offline anymore. But we also tested the rehearsal hypothesis and that predicts
that adolescents use online self-disclosure to practice, to rehearse, and they later start
to disclose offline. And we see it in this graph. This is a three-way longitudinal study in
which we measured both types of self-disclosure and what we see is only a path from online
self-disclosure at time one to offline, which is support for the rehearsal hypothesis because
if there was displacement hypothesis we would have seen negative relationships
overtime between going from online self-disclosure to offline. So what we see here is that adolescents
seem to use the internet to explore intimacy related issue with their close friends and
we also know that since 2.0 applications that adolescents communicate mostly with
their existing friends. In the 90’s when we had newsrooms and chat-rooms it was more that
they used the internet to talk to strangers but nowadays most adolescents use it to talk
to their friends. In another longitudinal study we investigated this effect of this online
self-disclosure on their friendship and what we found is that their online self-disclosure
was related to their perceptions of friendships at time two. So social media lead to higher
online self-disclosure, lead to higher perceptions of the quality of the closeness of their friends. The third social skill, sexual exploration, in one of our earlier studies we had found
that pre-adolescences nine to twelve years old experiment almost three times as
much with their identity than later adolescents. They most frequently pretended to be older,
more beautiful, sexier, and for the boys more macho. And you know looking for pornography
on the web has become normative behavior. in adolescences. Here are the most recent prevalences
and you see how often early and late adolescents deliberately search for sexual information
on the internet and you see they often do it together and they have great fun when they
do it and what you also see and you see boys and you often see here is how often they are
accidentally exposed to such information and remarkably you see that boys are deliberately
far more often expose themselves to pornography but remarkably boys also accidentally are
exposed more often than girls. We still don’t understand this later result.
Although exchanging sexual information between partners is normative behavior in adolescence
when you do it on the internet because of the distributability and the permanence,
it’s often risk behavior, and we have also investigated and you will see here a small
percentage of adolescences engage in sexting, sexual risk behavior, sent nude photos to each other,
talk about sexual topics, search for partners to have sex and here you see this is based
on a four wave panel study and we used accelerated cohort sequential design with h and the time
variable and you see here this is online sexual risk behavior. You see there’s a peak at around fifteen
and sixteen years and if you compare that with offline sexual risk behavior you see
another path. These are the high risk takers. You see that it starts later and we know that
the average age that teenagers have sexual intercourse is around seventeen and they start
with online sexual risk behavior earlier than offline sexual risk behavior. We’ve
done far more studies than I can show now that indicate that adolescents change because
of their own communicative behavior. We have mostly found positive behavior expression-effects
and these are related to self-esteem, their social competence, their friendship initiation,
their perceived friendship quality, and well-being. However we also found that five to ten percent
of adolescents are bullied online. We also found that five percent engage in sexual risk
taking, that seven percent get mainly negative reactions in their online profile, and that
five percent of adolescents are compulsive gamers of compulsive social media users and
we have also found negative effects, small negative effects, of their exposure to pornography,
on their attitudes about sex, and about women. And you know these small relationships, they
suggest that there are in this case also huge individual differences in susceptibility.
Research into expression-effects has hardly investigated individual differences and this
is in my idea an important task for future research. Thanks for your attention you can
follow our publication. We published everything online. And if you’re interested, everything
I told can be found on our website. thanks for your attention. My name is Marion Underwood. The title of my talk will be The BlackBerry:
Capturing the hidden World of Adolescents’ Digital Communication. I bring the perspective
of a child clinical psychologist who has always been fascinated by aggressive behavior and
as someone who studies aggression I’ve had long standing distrust of self-report measures.
So in my observational work of aggression behavior I have always tried to watch aggressive
behavior and I got into studying digital communication because students in my longitudinal study
were coming into my lab clutching cell phones and texting. I was asking them
to interact in front of cameras and they were, much to their parents mortification, desperate
to text on their cell phones. And I realized, what if I just captured their digital communication. It would be so much of a window into their social world. It’s been an honor and an adventure
to see the content of adolescents’ text messaging. Facebook and now Twitter and Instagram also.
I really thank the families and the children who participated in this longitudinal study
for nine solid years. I am also very grateful to the National Institutes of Health for their
ongoing support for this work. So the pioneers studying internet communication realized early
on that online communication is a real window into the world of the adolescent peer culture
and that the internet could be a context onto which adolescents project normative developmental
issues such as identity development and sexuality. In recent years adolescents have become heavily
engaged in a different type of the digital communication – text messaging. All of these
statistics, almost all of them, are from the Pew Foundation. These are surveys conducted
in the United States. Text messaging likely provides really important developmental opportunities
for close communication with peers, micro-social planning, literally saying “I’m at the movies,
where are you?”, figuring out where to meet, communicating about school work and exchanging all kinds
of information. This is too many statistics to read but let me just draw your attention
to a couple. Fifty-four percent of adolescents contact friends daily via text messaging. When
you ask adolescents how they prefer to communicate with friends, they say they prefer text messaging
over all other forms of communication including face-to-face interaction. In one study fifty-four
percent of girls and forty percent of boys reported that their social lives would end
or be greatly worsened if they could not text. Rich Ling has pointed out that texting
is probably a life phase phenomenon, a practice prized in adolescence because it is inexpensive,
it is discreet to the point of being almost subversive, it can be done in many situations
in which cell phone calling and internet communication are not possible, it is more private from adults,
it is easier for parents to spy on your Facebook messaging than it is for them to spy on your
text messaging, it’s more private and it’s a form in which you can play with slang and
develop your own language for interaction. Many people have maligned adolescents’ digital
communication and people are of sort of two extremes. They think it’s silly and not worth studying
or they think it’s terrible and ruining interpersonal relationships. You hear a lot of talk of adolescents
being addicted to their phones, addicted to digital communication, and I think it’s important
to stand back and to realize that young people are intensely engaged in this form of communication
because it meets their basic development. needs It’s a way to explore basic developmental
issues such as sexuality and identity and it’s a way to seek social support like that.
It’s also a way to engage with an imaginary audience and this is an old idea put forward
by David Elkind, that adolescents go around the world assuming that everybody is watching
them all the time. Social media is the imaginary audience come to life. Anything you post can
be immediately seen by your hundreds if not thousands of friends or followers. Marwick and
Boyd have pointed out that teens use the affordances of social media to gather attention, to involve
themselves in others’ lives and to manipulate public perceptions of who they are. I think
the third reason why adolescents are so intensely engaged in digital communication and maybe
why all of us are is that these devices and the type of communication they deliver to
us are powerfully reinforcing and they operate on an intermittent reinforcement schedule.
So this is my iPhone. I call it my iPhone Senior. It’s the big one. It’s big enough that I can
see it and I love it. I don’t love it because it’s pretty or its cool. I love it because
every once in awhile when I push the button and I check my email or my text messages, something
incredibly wonderful comes through. Now a lot of the time I check and it’s nothing. It’s
work email. It’s chores I have to do. It’s junk. It’s something I want to ignore. But every
once in awhile I push that button and it’s a text from my college daughter or from my
high school daughter or it’s an email that a paper has been accepted. These deliver to
us reinforcement on a powerfully reinforcing intermittent schedule and that’s just as true
for the adolescents as it is for adults. A lot of research on all forms of digital communication
has relied heavily on self-report measures and this has been pioneering work. But as I
said my training has always been to observe behavior. So what I want to do today is to tell you
about our research in which we’ve used technology to capture the actual content of adolescents’
text messaging and social media communication. I’ll focus on text messaging today just
because my colleagues are focusing more on social media, but we are capturing social media
also, so I want to tell you about the BlackBerry project, and this was the later part of a longitudinal
study that started in 2003 when the students were nine-years old and in the third grade
and members of this ongoing study were provided with BlackBerry smartphones in the summer
before they started ninth grade. Fortunately for us this was in 2008 when BlackBerries
were still really cool and iPhones were not required. BlackBerry enterprise servers allowed
you to archive or allowed services to archive digital communication. We were able to save
all of their text messaging, email, and instant messaging to a secure online archive as they
moved through grades nine through twelve. and Actually we kept collecting all of their digital
communication through their second year past high school until about age twenty and I will
say that all students and parents in our study knew that their electronic communication was
being captured and coded. There was full consent given every time we saw them. There was no
secret about this. We did not limit them to our devices but we think they use them really
heavily because usage of the study-provided BlackBerries was high. We had a mean of fifty-five
text messages sent and fifty-five received per-day. There were absolutely no gender differences
in frequency of texting. There was high variation. One male member of our sample regularly sent
and received over 19,000 text messages per-month for several months and on average over half
a million text messages poured into our archive each month. These rates of usage map very nicely
onto what the Pew people find when they do their surveys. So people often wonder given
that the students in your study knew that you were archiving all of their text messaging
were they communicating openly. I believe they were. One way that we looked at this was just
to take a two day sample of text messaging gathered in the fall of 2009 when they were
in the tenth grade. This included 43,505 text messages. We ran it through LIWC, which is one
of the ways we’re coding our data. LIWC stands for Linguistic Inquiry and Word Count, developed
by Jamie Pennebaker at UT Austin. LIWC counts the linguistic properties of speech. So in
our two day sample we found that about six point six percent of utterances contained sexual
themes and seven percent contained profane language and this is comparable to what had
been found in unmonitored chat-rooms in some earlier research. Much of the communication
was positive and neutral and I’ll say more about that in a minute. But some was very frank
and profane. If they were monitoring, if they were censoring themselves, I don’t know what
they would have said otherwise, and I’m not reading this out-loud but you get the idea.
Some of it made me want to wash my eyes. So the truth is that most text messaging is positive
or neutral. We coded our text messaging data in two ways. We did micro-coding, very careful
line by line coding, of two two day samples per-year in grades nine through twelve and
we have also developed filters to use LIWC to count the linguistic properties of the
text messaging. Of course LIWC is dictionary based and so we need a filter to fix all the
spelling. The truth is our micro-coding reveals that seventy percent
of adolescents’ text messaging are positive or neutral and the proportion of their texts
that are positive or neutral predicts decreases in internalizing symptoms across the first
year of high school. It’s also important to say that adolescents use test messaging to
engage in intimate supportive behavior. I am a person that is interested in anti-social
behavior and bad behavior so that’s mostly what this talk is going to focus on. But there
was plenty of it that was positive and here’s an example of what I think is a really nice
text message. The boy says, “you’ve just got some problems but hey who doesn’t as long
as you have someone you can talk to and tell things to I promise you things will get a
lot better there are so many people in the world with problems but they refuse to tell
people and open up because honestly these days you can’t trust anyone but I’m promising
you I’m here to hear it all and everything is between us two wait what does your mom
do while all this drama between you and your dad going on and did your parents send you
to live with them or were you just super close to them growing up?” So for all the misspelling, for all the informal communication, this is actually a really nice supportive text message
and we saw a lot of this in our archive. And still, although most of it is either positive
or neutral, specific negative types of communication seemed to predict maladjustment even if their
base rates are low. So what I want to do now is to tell you a couple of brief examples
of developmental questions that we’ve tested using these data. Let me be clear, we have only
scratched the surface. We have an enormous amount of data that we’re working with and
like I said this was a longitudinal study so we saw everybody in person every year: the
child, a friend, the parent. We also got teacher data on the child’s school academic and social
adjustment at school so we have a lot of information about these young people’s offline lives as
well as their online communication. The first study I’ll describe is a study to see
whether text messaging about antisocial activities would predict increases in rule breaking and
aggressive behavior across the ninth grade year. My outstanding former student Sam
Ehrenreich wanted to know whether adolescents engage in antisocial communication via text
messaging, how much do they text about antisocial behaviors, and he wondered whether texting
could be a context for deviant peer communication, deviancy training, in the way that Tom Dishion
and his colleagues discuss it. He also wondered whether texting about antisocial behavior
would lead to increases in externalizing behaviors. I’ll just give you a short overview of
the data collection because it helps you understand the different sources of information that
we were working with. The data I’ll describe for this study were collected from the spring
of the eight grade year when we got the baseline teacher ratings of externalizing behavior,
then summer vacation — we got participant and parent reports of externalizing behavior. And
then over the course of the school year we coded two two day samples of text messaging:
one in the fall around homecoming, which is a special Texas tradition with a big football
game and a special dance; it’s the social event of the season, and then two days around Valentine’s
day, and then we collected at the very end of the school year teacher outcome
data from the ninth grade teachers, and then over the following summer vacation we got
parent and participant outcome data. We were measuring externalizing behaviors using the
child behavior checklist measures, the Parent CBCL, the TRF, the Teacher Report Form of the
CBCL, and a youth self-report form. So that’s kind of the overview. Like I said, we coded
four days of SMS communication and our coding system was designed to primarily capture social
aggression as well as other key forms of communication. But we carefully coded antisocial behaviors
as well and these are the ones we examined: property crimes, physical aggression, substance
use, and rule breaking. It was not difficult to code these types of text messages reliably.
The results showed that first of all adolescents did communicate about antisocial behaviors
in text messaging. You can see here that about fifty-nine percent of our adolescents engaged
in some type of antisocial text messaging and of those who did their mean number of
texts over four days was seven. You can see the percentages for the type of antisocial
texting. But what we take from this is that text messaging about antisocial behavior does
occur. To give you a flavor of what these communications are like here is an example. I apologize for
the size of this type. I will read this. This is Michelle talking to Griffin and the
only thing we’ve changed here is names and a few details to de-identify it. Michelle
says to Griffin, “Have you seen the Pixar movie Ratatouille while flying high? You need to”
and then Griffin says to Michelle “I should but I don’t feel like flying high right now
I’m waiting until the mall tonight” and Michelle says to Griffin “wow dude how many wings” which
means how much marijuana do you have. “it’s like a freaking endless supply” and Griffin
says to Michelle “LOL I have bought eighty-five dollars of weed in all my weed life it lasted
a long time still lasting.” Michelle says to Griffin “dang son”. Griffin says to Michelle
“I didn’t buy it all at once and I use a pipe which conserves weed it’s amazing what pipes
can do”. Michelle says to Griffin “yeah I see that now” and he says “yeah I always get my
stuff from my guy in town he gets me lot for cheep it’s like twenty dollars for five
grams”. Michelle says to Griffin “aha you’ve got the connections” and Griffin “yeah I only
buy from Nick that is A1 stuff that’s the best stuff ever” and Michelle says to Griffin
“oh yeah well you know I don’t want to get hooked” and Griffin says to Michelle “LOL don’t
it’s expensive but yeah don’t buy from Nick it’s too expensive and it would ruin other
types of weed because they won’t get you as high as A1”. So what you see in this example
is deviancy training at its finest. You see encouragement and establishing the behavior
as normative. “Hey have you seen the movie Ratatouille while flying high” you know like as one does
. And you’ve got all the connections. You see instruction: “and I use a pipe which conserves
weed”; he’s teaching her you’ve got to get this special pipe. It’ll save you money. “my guy gets
me lots for cheep like twenty dollars for five grams” and so on. So what we wanted to
see was whether this type of antisocial text messaging would predict higher rates of involvement
in antisocial behavior prior to entering the tenth grade. These results are regression
models that include gender, baseline rule breaking, so this is rule breaking at the start of the
ninth grade year, total utterances is just the frequency of text messaging, we wanted
to know that the antisocial content of text messaging predicted increases in antisocial
behavior above and beyond frequency so total utterance is just frequency, and then
antisocial utterances number of text messages or percentage of text messages that were
antisocial. And what you can see is that antisocial text message utterances predicted increases
in rule breaking behavior across the ninth grade year according to self-report teacher
report and parent report. Similar results were found for aggressive behavior except that
the results weren’t significant for parent report but for self-report and teacher report
text messaging about antisocial behavior predicted increases in aggression according to self-report
and teacher report above and beyond gender, above and beyond baseline aggressive behavior,
and above and beyond simple text message frequency. So we think what this shows is that adolescents
use text messaging as a form for antisocial communication and that antisocial talk via
SMS significantly predicts later involvement in rule breaking and aggressive behavior. Text
messaging really is a meaningful venue for social interaction and can be a significant
predictor of later behavior. We dismiss this form of communication at our own risk. We also
think a lot of the content of our archives suggests that parents and school administrators
need to redouble their efforts to monitor SMS in school. There is a lot of texting in
school and a lot of it is about where to meet and what bathroom to smoke in, what kinds of substances.
I want to tell you about another example where we looked at a particular form of text messaging,
one that again has a fairly low base rate but is predictive of psychopathology. This
study is called Let’s Talk about Sex: Sexting Early Sexual Activity and Borderline Personality
Features in Adolescence. The purpose of this study was to move beyond sexting in the context
of sexually explicit pictures and to identify the prevalence of sexting defined as sending
written messages with sexual content. The technology we used does allow us to capture the pictures.
We have the pictures. We never opened them up. We never plan to look at them, don’t want
to, never will. There was a lot of writing about sex in the text messages and that’s how we
define sexting for the purpose of this study. We wanted to see whether text messaging about
sex related to increased sexual activity and we also wanted to see whether text messaging
about sex would predict Borderline Personality Features above and beyond sexual activity.
Other goals were to identify the prevalence of sexting in the typically developing sample,
and like I said to define it not as sending pictures but as written messages. So here’s
an example of a sexting conversation in which adolescents are discussing what sexting is.
Harry says to Sally, “awkward question have you ever sexted before” and Harry says to Sally
“what does that even mean” and she says “you don’t understand what sexting is?” and he says
“not really” and she says “oh it’s like describing what you would do to the other person during
sex” and then Harry says to Sally “oh so I don’t know what to say now” and then Harry says
to Sally” oh I thought it was like sending pics or something but I wasn’t sure. have you
ever done it before” Now I could show you lots of profane sexting exchanges and I’m not going
to do that but what I like is they’re actually talking about what it means. They’re not sure
and the girl is saying “oh no it’s not just about pictures. you like write stuff about
what you would do” and we saw a lot of that in our archive. So the method for this study
is pretty similar to the one I showed you before except we were looking across the tenth
grade year predicting to twelve grade. We measured Borderline Personality features using
the McLean Screening measure for Borderline Personality that was given prior to entering
the tenth grade and then again following the twelfth grade year, so that’s ages approximately
sixteen and eighteen, and we used a combination of questionnaires to assess risky sexual activity
adapted from some existing surveys and then across the tenth grade year we coded two two
day transcripts for discussions of sexual behavior. Like I said we coded these two two
day samples from both discussions of hypothetical and actual behavior. So let me show you just
some descriptive statistics. First of all you can see that about sixty-five percent of our
sample was involved in text messaging with sexual content. The rates of texting about
hypothetical sexual activity were higher than the rates of texting about actual sexual activity.
Some of the texting about actual sexual activity seemed like a lot of talk. It seemed like they
couldn’t possibly be doing what they said they were doing while they were texting. But
anyway, actual and hypothetical talk about texting about sex were highly correlated.
About 118 of our participants sent or received a total of about 3,594 sex talk utterances.
Boys were more likely to send texts about actual sexual behavior but there were no differences
in texting about hypothetical sexual behavior. Just for the sake of time I will say that
our first set of hypotheses about sexting predicting sexual activity were all supported.
It was absolutely the case that text messaging about sex raised the likelihood of an early
sexual debut of various types of risky sex and just as one example of one of these findings was
adolescents who engaged in sex talk and texting were one point four three times more likely to have
had sex by age eighteen. Texting about sex at age sixteen predicted the likelihood of
an early sexual debut and other forms of risky sex behavior. We also thought
it would be interesting to see whether engaging in sex talk in text messaging would be positively
related to borderline personality features at age eighteen and would predict growth in
those features from age sixteen to eighteen. Borderline Personality Disorder is a DSM5
diagnosis that’s categorized by a lot of instability, impulsive behavior, unstable chaotic interpersonal
relationships, and desperate attempts to avoid abandonment, among other symptoms. We believe
that text messaging might be one context in which adolescents perhaps at risk for Borderline
Personality Disorder might express their desperate need for attention and for connectedness. So
again, just going quickly for the sake of time, this is just an example of one of our findings.
What we wanted to know was whether text messaging about sex predicted increases in Borderline
features above and beyond baseline Borderline features, above and beyond gender, and above
and beyond our sexual risk variable. Since we know sexting predicts to sexual activity
we wanted to know even above and beyond the risky sex behaviors was text messaging about
sex predicting a Borderline Personality features at age eighteen. And the answer is it was and
for every model we tested, for every different type of risky sex included sex talk texting
them about sex in grade ten predicted increases in Borderline Personality features and there
were no gender differences in these findings. We believe that this suggests the constant
availability of text messaging might be problematic for some adolescents, that sexting might facilitate
the impulsive sensation seeking features of Borderline Personality and that in addition
sexting might create added pressure to engage in sexual activity because the expectation
has been set. I worry that there’s something about sexting that may exacerbate Borderline
Personality features. We’ve only started to examine how text messaging relates to quality
of relationships and this is just a very quick look at this. We looked at the correlations
between some of our texting variables and qualities of friendships, self-reporting qualities
of friendship, on a friendship qualities measure in the ninth grade and the very sad side of
the story is that there was no relation between anything about texting and any of the positive
features. But you can see that there were relations between text messaging variables and negative
features of friendship and that’s worrisome. We want to understand a lot more about what
qualities of text messaging predict what qualities of relationships, and that means that we need
to code our data in a new way and need to look at some variables that we haven’t looked
at yet. We are examining social media. A lot of this has been covered already. I will just
say that right now I’m doing a study in collaboration with the Anderson Cooper 360 program and Bob
Ferris at UC Davis where we’re looking at the social media communication of thirteen
year-olds, and thirteen year-olds love Twitter. It’s the preferred platform. 76% of them. I’m
sorry Instagram. And the thing we have to remember about social media is
that they’re not just posting but they’re reading these enormous feeds of all of their
friends and followers and so they’re exposed to all kinds of things on those feeds all
the time. So what I believe is that adolescents use social media and text messaging to meet
their developmental needs and basic development theories can be tested in this context. We
need to understand much more about how digital communication affects the quality of adolescents’
relationships. In the words of Dana Boyd, adolescents are not addicted to social media. If anything,
they are addicted to each other. Adolescents are living their social lives online and we
need to move toward understanding digital communication not as a distraction from the
real relationships but as a significant part of their social worlds. Thank you very much. I’m in the school of information at the University of Michigan. I schools for those of you who
aren’t familiar come out of the library tradition so their kind of like librarians-discover-
the-internet. My training was in communication theory and research so what I’m really interested
in is how these new online tools kind of reshape our interactions with one another and to the
extent we get value from these relationships. So today what I’d like to do is share some
thoughts about the definitional properties of social network sites and how they can support
social relationships and our ability to extract value and give back to those in our social
network. The overall argument that I’ll be making is that the social network sites
in particular have social and technical affordances that help individuals invest in and maintain
and extract value from their social networks. Obviously there’s a whole slew of social
media applications. Most of the research that I’ve done focuses specifically on Facebook.
Facebook is a social network site which is slightly different than some of the other
apps that you see represented up here. Facebook has about 1.4 billion active users and more than eighty
percent of those are outside North America. And I think Patty actually set come of
this up very nicely – this distinction between traditional mass communication and social
media. This is a page from the first
American newspaper. Newspapers are broadcast forms of communication, mass comm., one too
many, and interestingly enough early newspapers actually had more interactive features. So
as I said this is the first newspaper in the US, printed in 1690, and it was a four page
newspaper but actually only three pages had print on them. The fourth page was left blank
so that users could right thoughts and opinions on it before they gave it to the next person. Of course many newspapers in the print tradition don’t have
those forms of interaction, nor do mass media like radio or television. One of the things
that you might have worried about if you were living around this time was the plague, and
here’s a recent story that I first encountered on Facebook about the plague being actually
distributed by gerbils, not rats. I’m not alone in getting some of my news from Facebook. In
2013 Pew did a study that suggests that thirty percent of all US adults, not just online adults but
all US adults, get news on Facebook and twenty two percent think of Facebook as a good place
to get news. This is just one example of the interesting ways in which our distinctions
between newer and older forms of media are collapsing and our communication and consumption
patterns are really being reshaped. So in the case of Facebook and news, one of the things
that I know when one of my friend points to something that I might be more interested
in it, that they’ve kind of done some kind of curation work in directing my attention
to those particular stories, and in the case of the gerbil black plague story this might
have been something that I might have missed if it were not for Facebook. So this example
of exposure to novel information is one of the benefits we get from our social ties. Other
benefits we get from our networks include social and emotional support. So what properties
define a social network site? In 2007 I worked with danah Boyd to curate one of the first
collections of social network sites. This was published in the journal
of Computer Communication and as part of editing this special issue we wrote an overview of
the academic literature on social network sites, which was not much at the time, and we
also decided to forward a definition of social network sites because one of the things that
was happening was that people were using these terms social network site or social networking
site to mean a whole host of things, not using the same term to mean different things, different
terms to mean the same thing, and it was really kind of stymieing our ability as researchers
to synthesize across fields which I think is especially important for work in this area.
The three things we talked about in 2007 as being definitional components of social
network sites were a profile, an articulated list of connections. So in Facebook that would
be the friends list, and then the ability to traverse through those connections. Obviously
a lot has changed since 2007 in terms of both the social practices and the technologies
themselves. So last year in a chapter we revisited our definition to kind of take into account
some of the things that had been changed since 2007, and you’ll see that the profile is still
here and still very prominent. The only one difference is that now instead of profiles being more
static with just kind of user supplied content often times profiles now in social media are
co-created. So my friends post to my wall, my friends tag pictures of myself, there’s also
server supplied information — things I’m doing on the site. And this is in distinction
to earlier more static forms. And the other major difference is that we’re really calling
out the importance of the stream in Facebook. This is the news feed, this kind of aggregated
collection of everything that our contacts on the site are producing. I wanted to use
these three properties to kind of frame my talk, which will focus on each of those.
The first one I want to talk about is the profile. I think user profiles are
a fascinating context to study questions of identity, how we think about ourselves,
and when you look at the really really early work on profiles – I’m thinking for instance
Sherry Turkle’s work on MUDs in the 1990’s — there was a lot of talk of them being these
fantastical places to explore different components of our identity. You could go online and
pretend to be a different gender or interact with people in an online context where everyone
was pretending to be an animal, for instance, and you know of course different online contexts
have different understandings and expectations about how much people will kind of play with
fantasy and in many of the spaces that we look at now there’s not really this kind of
sense that what happens online is so different from what happens offline. There’s a sense
that online representations are meant to connect with our offline identity in some way. So thinking
about again profiles in 2007, as I said very static. This is mostly information that I’ve
created and more recently with Facebook at least we see many other kinds of identity
information, information provided by the server itself or by my friends on the site.
So I’ve done a little bit of work looking at self presentation in an online dating context
and I think it’s a particularly interesting context because you have the same affordances
that allow you to basically engage in deception or in kind of fantastical presentations of
identity but then you also have this expectation that you’re going to meet face-to-face
at some point and so you want to use the affordances of the site to put your best foot forward.
But you can’t lie in such a way so that when you meet someone for coffee they don’t get up and storm off because you look nothing like your photo. So thinking
about identity more globally, Rene Descartes struggled for months thinking about
identity and who am I and this question is pretty routinely posed to anyone who wants
to create a profile on an online dating site. So a few years ago I did some research with
colleagues and what we did was look at profile-based self-presentation. We collected the
profiles of match.com and other online dating site users and then we asked them to come
into the lab and when they came into the lab we asked them to get on a scale and to show
us their drivers license so that we could see their age and then we compared the two
what we found is most people lied on one of those three dimensions — height, weight, or age. But the magnitude of the deceptions was very very small, something that you might not really
notice if you met the person face-to-face. We also interviewed our participants and really
what we were curious about is their thoughts about what the profile represented and also
what kinds of discrepancies online and offline discrepancies they thought were acceptable
and to come up with a conceptual framework for how people think about the profile in
this particular setting. And what we found was that the asynchronous nature of profile
creation and consumption allowed people to basically select from a library of selves
when they created their profile. So the fact that I was writing a profile today but I don’t
really know when it will be read – maybe it’s not going to be read for six months, and maybe
in six months I’ll lose ten pounds therefore for me to list my weight as ten pounds lighter
it’s not really a lie. So this is kind of the way that people understood these profiles
and what they represented, so when it came to the acceptability of which of these discrepancies
was problematic, if the trait was something you could change, it’s okay to fib a
bit about your weight. It’s not okay to fib about whether you have children. One of those
things can be changed, and also the magnitude. So if it was a very very big difference it
was much more problematic than smaller ones. I’m not going to go too much into
this data collection but I just wanted to give you a sample of quotes. This is someone
who’s talking about their profile and what’s acceptable and what’s not and she’s basically
saying that if you have the intention of becoming that person it’s not as problematic.
In this case she’s saying I don’t go to the gym now but if I had a job where
I had money and I could pay for a gym membership than I would. Therefore it’s okay for me to
list that on my profile. We also found people incorporating aspects of their past selves. Here was someone who created the profile when they weren’t smoking, started smoking,
but then didn’t go back to correct it because they wanted to stop but just hadn’t stopped
yet. So the conceptual framework that we suggest is that people weren’t thinking about
their profile as an exact one-to-one replication of who they were offline at that moment but
rather it was a promise to this imagined audience of who they might be by
the time they met face-to-face. was I want to talk more
specifically about social network sites which is actually more of the work that I’ve done
recently, thinking about this articulated network of friends – what are some of the ways
in which we interact with these people? I’ll note that this is slightly different
from the other forms of social media. For instance on a blog I can have a
list of other blogs that I attend to but those people aren’t necessarily connecting back
to me. So the definitional characteristic here is the visual nature is both ways. On Twitter,
for instance, I can see who follows me, but also who I follow. We have these friends,
they’re also called contacts, connections followers, and these connections really serve multiple
purposes on a social network site. They can mark relationships, they can delineate
who can access what content, and they also serve as a network through which users can
browse profiles and discover friends in common. Now there’s much more emphasis on the social
graph, as opposed to just kind of the ego centered friends network. The social
graph is the entire collection of all networks on a particular site and media companies in
particular are very interested in leveraging the information embedded in social graphs. The idea is that if I’m friends with fifteen people. It may be that what those fifteen people
are interested in is also going to be interesting to me. As these sites develop these algorithms
for what they think we’re interested in they’re really relying on and then trying to leverage
information in the social graph. The key theoretical framework that my colleagues and I have used
to understand some of these social exchanges is social capital. This generally refers
to the benefits we get from our social relationships. This is an example of a barn raising. This
could be considered a form of social capital. It’s an established theoretical perspective
in the sociological literature although now many other disciplines are using it. In the sociological literature you would look at your position in a network, the size of your
network and what are the resources possessed by those in your network, and this helps us
understand how these resources can be deployed and activated via social relationships. One important component of social capital is reciprocity, this notion that if I help
someone I’m going to expect that they will help me in the future. Nan Lin, who does
work in this area, actually defines social capital as an investment in social relations
with expected returns in the marketplace. Most of the work that
I’ve done focuses on bridging social capital and we know this primarily as the strength
of weak ties. The idea here is that, due to homophile pressures, we usually share
a lot of characteristics with those we’re close to and we’re more likely to access new
world views and novel information from our weaker ties, our friends of friends. In early
work in 2007 using an undergraduate population we discovered that self-reported use of Facebook
was predictive of social capital bonding and bridging social capital. Why might this be? There’s a few things that we would point to. One is that through the
status update you can broadcast requests. You can ask for social support. You can ask for
information. And it’s very easy then for one’s networks to answer, to comment on these requests. Social network sites also facilitate exposure to friends of friends, which we know are more
likely to provide novel information and they also allow people to engage in relationship
maintenance and social grooming. Okay, what do I mean by social grooming? Think about
primates and grooming. This is literally the picking of nits off of one another. In
primates this helps develop bonds within groups and also creates the expectation that this
attention will be reciprocated. As humans we don’t engage in this kind of grooming very often but we do engage in social grooming. The idea here is that when you go to write
happy birthday on a friends page on Facebook you have communicated to that person “I care
about you. I’m paying attention to you.” okay When we like a picture, this is a way
to indicate attention. So one of the things that we’ve done in this study in JCMC that
was published last year is we used survey data from about 600 non-faculty staff at a
university in the United States and we introduced a scale that we call Facebook Relational Maintenance
Behaviors. These are questions that people would either agree or disagree with, and you
can see that items or things like when I see a friend or acquaintance asking a question
on Facebook that I know the answer to I try to respond. When
you see a need expressed by someone on Facebook, do you try to help out? And when we do a regression
predicting bridging social capital, you can see the social grooming behaviors. People who report that they try to act when they see a request from their networks, this
is very predictive of social capital. And why might this be? Well, I think that a few things
are going on. One is this expectation of reciprocal behavior. If you are more likely to
answer questions from your friends on Facebook you may have more of an expectation that they’ll
help you out when you ask for help. There’s also this social grooming function and signaling
attention to certain ties within the group. At least in Facebook, the news feed algorithm
tends to privilege information with people that you have past exchanges with. So interacting
with someone then may help you become more visible to one another within the news feed
itself. And then finally comments on friends updates and wall posts are seen by the friend’s
network, not your own. So think about the strength of weak ties. When I write a status update
it’s just going out to people that I already know. When I comment on someone else’s status
update it’s going out to their network and therefore I’m connecting to these friends
of friends. So why is it important to actually do something, to click, to write a comment? Think about the constraints of the system. Basically if you don’t take
some action and leave a visible trace, there’s no way for someone to know whether you saw
that post or not. So clicking in Facebook serves the same function in some ways as the whites
of our eyes in face-to-face interaction in where we would make eye contact with someone
or you could see where someone was looking. So in humans, the whites of their eyes are several times larger than other primates and really this
evolved because it allows for coordination. You can see where someone’s looking. You
can figure out what they may do. And the idea here is that in these mediated contexts we
don’t have eye gaze to indicate attention and therefore we need to take some visible
action that kind of serves that same purpose. Aand last but not least I want to talk
about the stream in Facebook. Essentially you have a status update that you can send out
and broadcast to those in your networks and then all of those updates are aggregated in
a media stream. In Twitter it’s the front page. This is the kind of collection
of content from your network, and we know now that most people spend a lot of time in the
stream itself. They’re not really clicking from profile to profile the way you were in
earlier versions of social network sites and so one of the things that we’ve decided to
do in this more recent work is to really focus in on the instance where people are asking
for help from their network as a kind of a really explicit window in which we can study
some of these social capital dynamics as opposed to asking kind of general self report measures
of you what are your behaviors online. So here you can see that social network sites enable
question asking and information sharing. This is something I posted a few years ago when
I was in Boston on sabbatical, asking my network what should I do with the kids before
we leave Boston, and I got a bunch of answers, some more useful than others. You can also
ask for social support. Here s a post from someone in my network who’s talking about
their beloved kitty who just died and you can see that they’re getting a stream of people
expressing sympathy. As I said, we’re focusing in on this recent work on how people are using
this particular status update field to ask for help and to get provisions of help.
All of these papers, I should say, are online so I’m not going to go into specific details,
but generally what we did was with collaboration with Facebook researchers we hand coded a
corpus of public status updates to look at what kinds of requests people were making
of one another. For instance, was it asking for commendation? Was it asking for a favor? Was it asking for informational help? The first thing we did was look at whether it
was asking for help or not. We called these resource mobilization requests. And if they
were asking for help, what kind of help are they asking for and also what was the cost
of that request? Was it something that could just be done online through a click, through
sharing something, or was it something that needed offline help, like help me move on Saturday. And what we found was that four percent of all public status updates were these mobilization
requests. Most of them were favor requests and polls or asking for information. Most
of these – seventy percent – were very low cost, something that could be done within that
online setting. In another paper in 2014 we looked at responses to those public mobilization requests – what was the comment stream like? And when we compacted mobilization requests
where people were asking for a response to those where people weren’t asking for a response,
what we found was that mobilization requests received more comments and also retrieved
those comments faster so suggesting that the audience did or the network did try to respond
when they saw people asking for some kind of support. And then of course what’s the
link to social capita? So in the third piece published in New Media and Society last year
we used the public hand coded status updates to train a classifier, a computer program that
we then were able to use to analyze private status updates as either being likely to be
mobilizations or not. And we were also able to invite users to participate in a survey
and what we do when we compare our self-reported social capital measures to these mobilization
attempts in these status updates. We found that those who posted one or more mobilization
attempts in this twenty-eight day period reported higher bridging in social capital than those
who didn’t post one. We also showed that our social grooming measures were positive significant
predictors of both bridging social capital, which is this exposure to information from
our weak ties, and also bounding social capital, which speaks more to emotional support and
is often associated with closer ties like friends and family. Finally, we found
that twenty-percent of the active users on the site posted one or more mobilization attempts
in the period. So it is more common than our earlier work suggested. In conclusion, social
network sites include a set of affordances that allow people to maintain these social
relationships and also to help people extract value from their connections. I believe strongly
that the connections we feel and the benefits we access via social media are just as authentic
and important as those sustained through face-to-face communication and I think we should stop talking
about offline interaction as being the real world and online interaction as being something
ersatz or fake. E-relationships are relationships. Thank you. Thank you to the visionaries at APS for having the foresight to put on an international conference. Today I’ll talk about or try to answer this
question: Can technology bring us true love? I want to begin by observing that having
a romantic relationship tends to be linked to all sorts of good physical and psychological
health outcomes. But the quality of your relationship matters a tremendous amount. Not all relationships
are going to have those positive effects. When we’re involved with somebody who’s compatible
with us that is when things are particularly good. The problem is that it’s not always simple
to meet somebody who’s compatible with you and for this reason in all cultures that we
know of there is some type of match making process. There’s been wide cultural variation
in how match making works but never until this century have we seen anything like how
match making predominately works today, which is through computers and intensive mathematical
match making algorithms. e-Harmony is the prototype here. They try to set you up with somebody who’s similar to you and what I want to talk about today is
the issues with the existing match making algorithms and I want to try to set up a vision
for what an effective match making algorithm might look like. I think nobody’s built
one yet. I’m certainly not going to suggest that I have one. But I want to push us in a
direction that I think holds some promise. Before I do that we need to ask this question: What
is compatibility? The reason why this question is so important to me is that Google, the big
tech company, held what they call a round table. They invited me out and a bunch of leaders
of the major online dating sites, but also human match makers, like old style match makers,
and the conversation in the room frequently revolved around men who are threes really
thinking that they deserved women who were nines. I actually made a fuss and I think became unpopular because I think that if the match
makers of the world think that way, the world is worse, because we don’t want to figure out
who are nines and who are the sevens and who are the fours and set them up so that they’re
not with anybody who’s better or worse than they are. We want something more sophisticated.
Imagine that here’s Dave. It is possible that there’s these other women
in the world (we’re going to talk about the heterosexual case in this particular talk).
What we want to figure out is in cases where we can find somebody who’s uniquely
compatible with Dave, not that he got the best woman in the pool and then the
next best guy got the next best gal in the pool. We want to find unique levels of
compatibility. And of course that is what all the match making algorithms are attempting
to do. How have existing match making algorithms, including e-Harmony and basically everything
that exists today going about building match making algorithms, and why are
my co-authors and I so skeptical about this. Well, it turns out that we can think of two
different types of data that you can incorporate into a match making algorithm. The first type
is individualistic data and any information that you can collect on individuals. The
typical sorts of information may include things like demographics and age and personality
and attitudes and values. I kid you not to say there are sites that are trying to match
people based on DNA. You can submit your cheek swab and they will match you up with
somebody who’s compatible with you based on your genetics, and if you believe that God
bless. This is the information that is available to all of the existing match making sites.
Now is it adequate? Well, a whole bunch of us, in fact I’ll list the names, a team of relationship
researchers, Harry Reis, myself, Benjamin Carney Sue Sprecher and Paul Eastwick worked very
hard trying to figure out what’s wrong with online dating in general. But one of the
major things we did was we tried to figure out, has anybody build a match making algorithm
that actually works. I did not start out skeptical about this question but what happened is with
fifty, a hundred years of relationship science expertise in
that group, we asked ourselves the question: could we do it? Could we take information about
personality information, about values or DNA or anything else like that and build an effective
match making algorithm? Not find the people who are good in relationships and the people
who are bad in relationships. You can assess things like neuroticism and that’ll give you
some predictive value. But the unique compatibility, the sort of thing I was showing you on the
previous slide. We decided that we ourselves cannot do it. And we know what the principals
are that the match making algorithms use because they’re very limited in terms of what kind
of information is available to them. If you have information about a whole bunch of men
and information about a whole bunch of women but the men have no idea that these
particular women exist and vice versa, you have individual level information. What can
you do? You can set people up who are similar you can set people up who are dissimilar but
the range of options you have for compatibility is actually relatively narrow. So the question
is this: Our intro to psych textbooks, our intro to social psych textbooks, our relationship
textbooks all tell us that similarity leads to attraction. Are they right? They’re wrong. Here’s what we know about similarity and attraction. Number one: homogamy is real. People
end up with other people who are similar to them. Is that sufficient to conclude that they
like people who are similar to them? It is not, because you meet people who tend to be
similar to you in all sorts of demographic attitudinal sorts of ways. Second: perceived
similarity. For those of you up on this literature, the sorts of paradigm that Don Bern innovated
in the early 60’s, perceived similarity, the belief that you have things in common with
a person – huge predictor of romantic attraction. Those pieces of information are of no value
whatsoever if you’re trying to build a match making algorithm. How can I report my perceived
similarity about all of the hypothetical women who might be in the pool? I don’t know how
similar I am to hypothetical entities that I don’t know exist. What they need to focus
on is actual similarity and it turns out that now we know a great deal that we didn’t know
even five years ago. It turns out if you look at actual similarity you get almost no attraction
whatsoever in terms of initial attraction. So my colleagues and I ran a set of speed
dating events, for example, and we looked at subjective similarity. If you think you’re
similar to the person you just met speed dating you’re way more attracted to that person.
But we had the Big Five. We had a whole bunch of values measures, social, sexual orientation. We had a whole series of different individual-level pieces of information. It collectively
predicts no variance in how attracted you are to people. Similarly, a team of scholars
has done a twenty-three thousand person, three-nation, nationally representative sample
looking at the Big Five and marital satisfaction and zero point
five percent of the variation in relationship satisfaction is due to similarity in personality.
Basically you’re not getting much attraction with this stuff at all. There is a second type
of data that in principal we can use and in practice nobody has ever figured out how to
use. This is couple-level data. You don’t want to know whether you have some similar quality
to another person in the abstract. What you want to know is when you got across from a
pint of beer or a cup of coffee together did the conversation sizzled. Did you feel
like, “Whoa I want to know this person better.” That’s dyadic level information. You
can’t get that from personality questionnaires or from cheek swabs. You want to know about
rapport. You want to know about the sizzle of the conversation, whether you laugh at the
same jokes at the same time, whether you have this type of connection. The problem is to
this point the individual level data don’t work. Is it even possible to
incorporate dyadic information? What would it even mean? What does it even look like to
try to build a match making algorithm that takes account of information that is actually
assessed at the couple level or the dyadic level. Now what I’ll do with this talk, I have
three different parts. The first part I want to talk about is some scientific developments
within social psychology that I think hold the kernel, this basic way of thinking
about compatibility that I think we might be able to leverage to build an effective
match making algorithm. After I talk about the science, I want to talk about new technological
developments, particularly in the domain of machine learning that I think we can use in
concert with the science to develop such an algorithm. And finally I’ll get very concrete with an example of what I think an implementation of this set of ideas might look like. So the
science is built on one idea. That idea basically is interdependence. If two people don’t
know each other than knowing any information about Dave tells you nothing about Alice. Knowing that Dave is happy or sad or his heart is beating fast or beating slow tells you
nothing about Alice. But as we become more attracted, as we even early on begin to develop
some type of connection together you start to get interdependence. You start to get just
a teeny bit of information that if you know that information about Dave you actually start
to have a little bit of information about Alice. There starts to be a real connection
and it is correlated with the extent to which the two of them are connected. The logic
behind this match making algorithm idea comes from basically two sets of findings from the
psychology literature. I’m going to breeze through this stuff relatively quickly but
one of the findings is if you have initial chemistry, particularly if you want this other
person to like you, you tend to engage in various forms of nonconscious synchrony. This can
be behavioral mimicry. It can be the prosody of your speech. It can be the way you’re using
language. But across a whole bunch of experimental studies that manipulate the extent to which
you would like this person to like you, what you see is the greater amount of desire for
the person to like you the more you nonconsciously tend to mimic that person’s various qualities,
behaviors, vocal style etc. Now independent line of research starts the other way. It says,
what if people mimicked you? What if you were non-consciously mimicked, like you were interacting with a confederate who was trained to mimic your body posture, mimic
your facial movements, etc. Would you like that person more? And the answer to that is
yes as well. These are two independent streams of research but I want to suggest that they
connect up in an interesting way. In particular what I want to suggest is that if two people like each other they will surely know that there’s some connection there. But
there’s some amount of the connection that is not accessible to their self reports but
that is accessible to their non-verbal. That’s the whole point of non-conscious synchrony
and if we can tap that non-conscious synchrony, if we can figure out a way to use technology
to tap non-conscious synchrony in real time, then we might be able to get a sense above
and beyond their subjective experiences of how attracted they are to each other and maybe
leverage that information, again above and beyond subjective experience, to predict longitudinal
or at least short term and perhaps longer term compatibility. Let me talk a little bit
about the science. I’m just going to give suggestive evidence for some of these ideas. This is some
research I’ve done in collaboration with a team of people including Molly Ireland on
language style matching. This is a 2011 Psych Science paper. Language style matching is
the way we structure our sentences, not the content of what we’re talking about. It’s not like we like people who tend to like soccer or tend to like tofu burgers. It’s not what you’re talking about. It’s the way you use language. They’re called function
words and they are meaningless out of context. There’s a whole range of words that
sticks together all the content words. Is the way we use these seemingly meaningless
words linked to the amount of attraction we feel to each other? This is a study that we ran at Northwestern University. Let’s imagine Dave again. The way
speed dating works I’m sure most of you know this already, at our events we had twelve men
and twelve women. These were heterosexual speed dating events. Dave would go on a date with
the first woman and I was the MC at these events I would blow the whistle. Dave would
rotate or Alice would rotate or something and basically it was the case that every male
met all of the females and all of the females met all of the males. So when we ran
this study we looked at match making. Specifically what we did was we transcribed and then coded
through LIWC, the Linguistic Inquiry and Word Count, the extent of language style matching.
Now let me tell you I would not be able to tell, if I were to read high language
style matched pairs and low language level match pairs, even after doing this research
I couldn’t tell you which was which. But the software knows. This stuff is very subtle and
what you can do is you want to look at percentage matching. Speed dating has this terrific quality, which
is at the end of the four minute speed date actually they did it that night. You indicate
with regard to each person you met do you want to see that person again. There is a very
good very objective measure of yes there is at least some spark here. She has to say yes
to him. He has to say yes to her. So what happens? Well if you have the low language style matching
pair you got about nine percent of the dyads matching. If you had high language style matching
pairs you got about thirty-three percent of the dyads matching. These effects control above
and beyond how much you reported liking each other, how much you reported being similar
to each other. This is tapping unique variants above and beyond what you’ve got going on
in your head. Now you might say, okay that’s four minutes, who cares? I actually
think that four minutes, if you can predict initial attraction, you’ve already got
a good head start on an algorithm. But in this second study we studied the instant messaging
logs of dating couples who had been dating on average sixteen months at the time that
they submitted their ten day instant messaging logs and we followed them to see who broke
up over the next several months. Turns out that if you are low in language style matching
you had about a fifty four percent chance of still being together. If you were high in
language style matching you had about a seventy-eight percent chance of still being together. These
effects still exist above and beyond any effects of self-reported satisfaction at baseline
and a litany of plausible alternative explanations. Now let me back up here. This is suggestive.
I’m not trying to be comprehensive here. but I’m just illustrating what I mean when I
talk about this idea that you can actually tap come sort of information with technology
that seems to be predictive of compatibility above and beyond people’s ability to tap that
stuff on their own. There’s a whole bunch of different research on body posture and so
forth. I’m going to fast forward and talk about the technology. Now I don’t know much about
machine learning. Some people in the room are probably knowledgeable about machine learning
but I studied up on it a little bit. Turns out one of the more basic forms of machine
learning is Siri Apple’s digital personal assistant and Google’s got one. Basically
you say, “Siri where is the conference site?” And Siri says, “Go this way, go that way.”
What Siri starts to do — and over time the technology is getting better — is it really starts to understand
speech in real time. Can Siri do this perfectly? No, but the rate of progress with regard to
voice recognition technology is staggering. If Siri works nearly perfectly, so to the degree
that within a year or two years or three years it works nearly perfectly, what can we do? We
can listen to a conversation with two people conversing, say through the internet, listen
to it in real time and have the text spit out, and I showed you on the previous slide
that once you have that text you run it through software and it can tell you language style
matching. And I also told you that language style matching seems to be pretty strongly
linked to initial romantic compatibility and relationship longevity. But Siri or this language or text stuff is I don’t think the most exciting way
to go. There are other ways to go. This is some research by Shichuan Du at Ohio State University, a computer science person, and what she’s interested
in is face recognition — literally this field called computer vision. What she’s done
is she’s identified ninety-four facial landmarks and what the computer can now do,
assuming that you have a camera on your computer (and frankly these days who doesn’t?), the computer
can start to distinguish happy from surprised. The computer, actually
starting recently, can distinguish those two from happily surprised. The computer can also
distinguish happy from disgust. And seriously this is the greatest thing that can ever happen
to you. If somebody looks at you with happy disgust, like yif ou told a lude
joke and she kind of liked it anyway, you’re in. So the computer can detect these sorts
of pieces of information. Now these pieces of information are useful in their own right.
If you can get people’s what Paul Ekman has called micro-expressions, people leak some
amount of their affective experiences on their face without realizing that they’re doing
it, you can use this information in a pretty basic way of trying to discern above and beyond
their sort of self-presentation or attempted self-presentation, what sort of emotional experiences
are they having, but more importantly you can look at the extent to which the two individuals
talking are altering their emotional experiences at the same time. What is the latency? How
immediately am I switching from neutral to smiling? And is that around the same thing?
And when I tell a joke and give a smile and you respond with happy disgust. We can
start to use these dyadic cues to get that sort of information about compatibility. Now, the last one that I’ll show you in terms of just illustrating the sorts of technology. This
is Bill Freeman from MIT and this is I think the coolest of the perspective things I’m
going to show you. I’m going to show you a video. On the pre-side to the left you’re
going to see the raw video input. It’s just a video of a man’s face but he’s not moving
but it is actually a video. It’s nine seconds of video. And on the right side I’m going to
show you what they are able to glean from that video once they sort of zoom in. What they’re going to do in this video is they’re going to take very invisible, just
minuscule changes in coloration as a function of heart rate like we would never be able
to detect, but they’re building the technology that can do it and what they’re able to do
is extract from video what your heart rate is. Watch this example. On the left is going
to be the live video, the actual feed. On the right is going to be the doctored feed. Now
that’s freaky, right? You wouldn’t want to do that. You certainly wouldn’t want to have a
speed date with your face doing that. But they can do that. What does that mean? That means
they know your heart rate. Not only do they know your heart rate, they know the heart rate
of the person you’re interacting with. They know the extent to which during the course
of your interaction your heart started beating as one. They know not only are you excited
or not that excited, they know about whether changes in your heart rate is diagnostic visa
vi the possibility of compatibility between the two of you. I could keep going in terms
of what machine learning is doing in technology but I’m going to start drawing to a close.
The third major part of the talk is this part about implementation. This part’s the easiest. This is technology. Basically think Skype or Google Hangout. There she is. That’s
the caller Alice and she is having an interaction with this guy, let’s say he’s Dave, and they
have a four minute speed date, a web based speed date. Here he is and he’s
talking to her and they’re chatting about it and the video is recording the sorts of
information I just talked about. The series type technology is getting the actual content
of what they’re saying, the video camera is starting to detect physiological sorts of
measures, facial coding expression, going to be able to incorporate that information too. Then that date ends and we move on to the next date, etc. Now imagine for a moment, and
this by the way I’m just making up, but imagine for the moment that we built a system
that goes like this. We have a series of web cam based speed dates. You sit at your computer
and have a series of these sorts of speed dates or formative interactions. Let’s say
after each interaction you then indicate your level of interest. Maybe you report on a three or four quick self-report questionnaires, of course the self-report matters too, I’m
not saying we jettison the self-report in favor of these subtler measures. We use the
self-report and also these machine learning type measures in order to get a fuller package.
Let’s say you do that Monday Tuesday Wednesday from 7:30 to 9:00. That means you committed
four and a half hours of your life to this. At the end
of those four and a half hours you have had brief interactions with sixty different people
in your community who are in the right sexual orientation for you and right age range for
you and you have information not only about how was the initial banter via the web, which
isn’t perfect but this is getting fairly close to face-to-face style if you’ve got the video
and live interaction, and you have all these non-verbal indicators that we can can use in our algorithm and start getting s sense not only the extent to which
you have this subjective feeling of attraction but also the extent to which you guys are
showing non-conscious syncretism across a range of different channels. And I don’t know
if this will work. Let me just tell you nobody’s ever tested this but
my hope, and I think I feel more confident than just hope, is that we might be able to
build and algorithm that at least incrementally maybe by five percent or ten percent or fifteen
percent increases above and beyond your self-reports the compatibility that you will have once
you get face-to-face with somebody – now doesn’t take a whole lot of incremental value to make
the world a lot better. What if we can use these algorithms to improve compatibility
by five percent right, we use the algorithm to improve compatibility by five percent and
to increase the likelihood by five percent that you really like the person you went on
a first date with? If you want to think about global happiness in the world, if it’s
true that these relationships are crucial, that the compatibility we can get things started
on the right track faster and therefore ideally make the world a better place. I thank you
for your attention I want to thank the speakers for giving me
an opportunity to comment in public on their talks I’ll try to use as little time as possible
because I know all of you have lots of better questions than I do I feel like a bit of a
voyeur here I have to confess that I got my first smart phone about six weeks ago I don’t
use Facebook I have no Twitter followers I’ve never online dated and I have Whatsapp on
my phone but I haven’t used it yet but I feel like a voyeur in another sense too and that
is and this was highlighted I think in Marion’s talk especially we get to look at what our
adolescents are doing and saying with each other at the very least in the general and
if you’re good at snooping you can maybe find out a little more when I was an adolescent
there was no way my parents were going to find out anything about what we were talking
about and in fact you know the conversation that Marion presented about dope I can tell
you I had that conversation but it was face-to-face and there’s no record of it so I suspect that
a lot of the kinds of things that people are talking about today are things that were going
on before except that they’re going on in speed now there’s a quote from C.S. Lewis
that I’ve always liked and that quote is isn’t it funny how day to day nothing changes but
when you look back everything is different I think he got it exactly backwards isn’t
funny how when we start looking at this stuff it seems that things are changing minute to
minute but when you step back and look at it maybe things aren’t quite so different
as they seem that is maybe some of the things that these folks are talking about and that
we’re seeing in adolescents in particular are the same sorts of issues and processes
that we’ve always been talking about but they’re on speed now clearly they’re happening faster
and of course we might ask ourselves why are all the innovations coming from adolescents
you know why is it that you know us older folks aren’t creating the innovations and
part of it is of course because adolescents are always more open minded to anything new
but another part of it is most adolescents are convinced that their parents are an epic
fail and this is going to be the new way of doing things and this is going to be so much
better and this is going to be so much awesome and this is going to be a way that we are
going to set ourselves apart because we communicate this way and it’s like totally different and
of course if we sit back a little bit we can just let them have their fun so for example
the examples that we saw about adolescent deviant behavior well these are all things
that we know have gone on for a long time they go on now they go faster they’re more
wide spread more people know about them they have more impact they have more lasting impact
so they’re more influential and we can talk about them being more potent in a lot of different
ways but in many senses of the word they’re simply the same things except accelerated
or if we think about the kind of processes that Patti was talking about self presentation
well that’s certainly not a new invention nor is self disclosure nor is sexual exploration
nor is cultivating social resources in fact when reading the preliminary versions of their
slides I started thinking about Catcher in the Rye many of the things that Holden Caulfield
does are things we now talk about on the internet and of course Holden Caulfield didn’t have
the internet available to him if we think about the kinds of social networks that Nicole
was talking about well you know adolescences is expressly about wanting to know what everybody
in your network is doing I remember as a kid sitting there and making charts of who was
talking to whom I didn’t know I was preparing myself for a career in social psychology you
know mobilizing friends maintenance these are certainly the kinds of things that we’ve
always wanted to do and I think it’s kind of interesting that you focus on weak ties
because thinking about it adolescences is about weak ties I mean you know there’s some
stuff about intimate relationships but it’s really about these weak networks that we have
and as we get older of course our network shrinks so all think is kind of a good example
of how you can use the social networks to develop some of the major themes of childhood
or the examples that Eli was talking about you know I really get a chuckle at what all
these online dating firms are actually doing because I remember filling out one of those
questionnaires as an undergraduate and that goes back to the 1960’s now they didn’t have
as good of ways of processing them and the computer took about six days I think to analyze
my profile and I don’t think that’s just because I was weird the algorithms these folks use
as best we know and I think we know pretty well they’re embarrassingly immature algorithms
to be honest they’re stuff that you could go into any social psychology class and construct
in a half hour and they cloak them in fancy language but they haven’t yet had the gumption
to try to test them in much more sophisticated minim I think some of the examples that Eli
presented I mean who knows if they’ll work but you know when I teach relationships to
undergraduates there’s always one part of the class that I’m embarrassed by and the
embarrassment is when they ask the question whose going to be attracted to whom and I
have to say I don’t know we don’t know sixty years of this work yeah we can narrow it down
a little bit if you’re a religious fundamentalist you’re not going to be attracted to an atheist
you know if you’re twenty years old you’re probably not going to be attracted to a sixty
year old you know we can cut it down a little bit but you know that’s not very impressive
beyond that so what Eli is describing is really short circuiting the questions saying I don’t
know who’s going to have chemistry with whom let’s let people interact in a meaningful
way with as many people as possible and see what emerges and who knows maybe someday we
will find out that people who have bigger muscles over here are more attracted to each
other and maybe DNA swabs will work although I kind of doubt it in another sense though
I feel like a voyeur in all this because everything that we’re talking about today I think is
great and in ten years it’s going to be ancient history and that’s the really sort of depressing
part of it you sort of have my pity because your research careers are probably going to
be dead in ten years unless you find something new you know this kind of technology it’s
a radianth I’m part of a group at NIOS whose exploring some of these let me tell you a
few things that are not just on the horizon but they already exist and we just have to
kind of integrate them devices that allow you to hold hands with your partner anywhere
in the world this is something called a Frebble by the way you can look it up devices that
you can put in your pillow that not only allow you to converse with a partner who might be
on the other side of the world but will also vibrate to the other person’s heartbeat potential
even regulate temperature the ability to interact through virtual reality with another person
who’s on the other side of the world or perhaps even something artificial including the kinds
of interaction that I don’t want to mention in public apps that allow you to have instance
visual communication anywhere in the world with anyone and in particular apps that allow
you to perspective take so that you can actually see the world the other person is experiencing
and of course I’m talking about Google glass which I think was a pretty dismal failure
not because I think the technology is a bad idea but I think they didn’t have a good model
for figuring out how to integrate it in people’s lives so these are the kinds of things that
are one the horizon there’s a wonderful quote from E.O. Wilson who said humanity’s real
problem is the following we have god like technology medieval institutions and Paleolithic
emotions so the question is not what can we do in technology the question is how does
it get integrated into human social life and what’s interesting about that is that you
then have to fall back on the same kinds of questions that we have about social life already
in what ways can social media compliment existing interactions as opposed to substitute and
I think Nicole touch on this in a very actually I loved your mentioning that it’s not real
and not real interaction they’re obviously just as real at what point does online interaction
become in authentic think about the movie Her for a moment of which I assume at least
some of the people in the room have seen that kind of creation at some point became dysfunctional
I think they kind of jazzed it up in the wrong way in the movie because they made her character
get weird and that’s when he started to have trouble with it but imagine that you had a
presence such as Her in your life at what point would you get tired of it initially
we’d all probably think it was fun but at what point would you actually get tired of
it could you even tell that it wasn’t a real person you know the Turing Test for androids
to what extent can soil media actually complement other kinds of relationships as opposed to
being the primary basis for people’s relationships I think we don’t really have answers to these
kinds of questions yet but I think that the way to address them is going to be to take
the kind of perspectives that we’ve heard today and combine them with many of the kinds
of models of social behavior that we’ve had all along and of course the beauty of this
is that that will require interdisciplinary research research that takes academics and
social scientists who are generally interested in these processes and technology entrepreneurs
now one of the problems here of course is the problem of money and so you might want
to ask yourself the question why are we so fascinated with these kinds of technology
if you follow CNN there’s some new technology everyday and lots of people click on them
why are we so interested in them why don’t we just sort of turn our nose at them I think
there are a couple of answers to that I think one answer is the novelty of it is intrinsically
self expanding there are many theories around about how people want to grow as human beings
how we want to expand our resources and our perspectives the novelty in these forms of
interactions is exciting it’s really terrific to be able to do this kind of thing and we
feel like we’re growing in a certain way and I think that we want to explore by developing
these new methods but that of course requires that they stay new the second thing I think
that there’s something about the hyper-drive nature of these forms of connecting with other
people that’s highly diagnostic of the human condition for us and about our own personal
circumstances they tell us something about what’s important to us what we care about
how we want to connect to other people how we want to meet our needs and I think this
is really very fundamental and interesting question about the human condition how we
can find out so much more about what’s meaningful to us the idea that Eli mentioned the idea
of having sixty speed dates while sitting home and I’d be in my pajamas and then I’d
use some kind of technology to cloak that you know to be able to do that would be really
wonderful but to the extent that entrepreneurs are motivated by money as opposed to the inters
value of this then I think there’s going to be a gap between how the science develops
and how it might be used to really better the human condition because we’ve seen in
many instances where the drive to patent things and the drive to control the technology can
actually interfere with the natural implication of it and I personally think that’s why Google
Glass died it’ll be back but in a different form so I guess this would be a good point
to open this up to comments from other people and thank you for your attention yes hi is
this working yeah hi my question is for Eli in your study with the speed dating and the
language models did you control for native speakers and whether or not those speakers
were multilingual or bilingual excellent question in general they were Northwestern University
undergraduates so most of them they were all fluent in English most of them were native
speakers I agree that sort of with a larger sample if you were to scale this up you would
need to account for those like cultural differences and the way syntactic structure get stitched
together but we don’t know yet that’s a good question yeah so because you’re discussion
algorithms and specific precocity when you have more than one language in your brain
it effects your precocity and it effects also your use of function words and your use of
content words so it’s something I think you’d want to control for at least in the future
thank you thank you thank you for your wonderful presentations
and I have one question the kind of underlying thought is that the internet changes the nature
of interactions or relationships I wonder if it’s not like we just keep interacting
but the nature of internet allows us to analyze all the data in a really great way we actually
have access to the data with access to the data virtual and all of that so now we have
actually better view but that doesn’t change underlying way of which we are interacting
this is actually a question for all of you I’ll start with adolescents I think in many
ways it’s the same and in many way’s it’s different so if you think about adolescents
have always been very concerned what other think of them concerned about how others view
them how other estimate them that they’re struggling to understand their place in the
social world if you think about social media it’s sort of like rocket fuel for those concerns
you post something online and you can immediately see how many people like it how many people
comment not only are you watching how your own posts are received you’re see how everybody
else’s are so you’re constantly reading this giant read of who’s doing what with whom and
your friends’ posts get way more likes and comments than yours do and what does that
mean about your place in the social world so I think it’s very similar the content of
what is said is similar perhaps but the audience it reaches is different in the old days we
could not send everything to a thousand people with the push of a button and I think there’s
more of a chance to gauge how it is received in terms of likes retweets follows whatever
the platform is and I think that means a lot at least to adolescents so that’s at least
one type of answer but my colleagues probably have other thoughts I fully agree and I also
full agree with Nicole you know especially in our research you know we as adults we keep
talking about the on and offline world but for adolescents the on and offline world is
completely intermingled there is no online and offline world and it’s also that they
mostly communicate with their existing friends true social media that’s probably the same
and they primarily use the internet for communication purposes but so what we see is that all the
problems the social problems in adolescences bullying love sexuality they are now being
communicated online and that has changed but the problems and the issues are still the
same hi thank you so much for such an interesting panel and I really enjoyed all of your presentations
and I had to give so much thought it’s really hard to come up with one question so I have
this kind of perspective future related question so from everything what we know for today
about research and social media and relationships from present studies what do we know what
are if summarized them all what would be the positive effects of social media on the relationships
and what are the negative effects of social media on relationships and then if you can
see the future maybe your idea so how in twenty years from now or more how then social media
will transform human relationships so if you have any future vision of that that would
be interesting to hear thank you wow that’s quite a so I guess I mean I think I think
the overall theme that I’m hearing and that Harry mentioned I think that the enduring
needs that we have as humans are consistent and we are now looking to new ways of meeting
those goals through the affordances of CMC I know Patti and I both looked to Joe Walther
for some of his really important work in kind of thinking about how CMC is different from
face-to-face it’s not better it’s not worse it has a different set of affordances a synchronicity
and I think if we look to some of the really really early work on relationships and interpersonal
communication and CMC the guiding assumption was that face-to-face communication was the
gold standard and everything else was somehow less rich or impoverished and now I think
it’s clear that you know we have a different set of cues from online interactions so maybe
I can’t see eye gaze but maybe when I look at someone’s online dating profile I can see
things that wouldn’t be apparent face-to-face I can see whether they know the difference
between there their and they’re right and if they use incorrectly I can make attributions
about their education the extent to which they’ve that you know how much time they’re
putting into this process so I think in terms of your question of kind of how are things
changing of what will the future be there’s a great article that looks at the difference
between being there and then beyond being there it’s Holland and Stornetta I want to
say and the idea is that you know for a long time we focused on trying to create technologies
that give us this same sensation of sharing time and space and another way to go about
doing this would be not to try to replicate face-to-face but rather to use it and use
these tools in new ways and I think some of the work for instance that Eli is talking
about with being able to use computer algorithms and video data right and on the fly processing
that we would look to to try to kind of capitalize on those affordances but not necessarily try
to replicate face-to-face I think that’s not I think that’s not going to happen and I don’t
think it’s as important as some of us seem to think it is I think one of the challenges
in the future is abundance how can we deal with abundance and I think you know and when
we talk about raising children when we talk about teaching children self-control is one
of the most important me parenting goals because you know all the technology is closer to our
skin it was on the PC on the desktop and it moved to our bag the laptop and not it’s in
our pocket and you know and all these apps they scream for attention and I think it’s
very you know I see many many adolescents struggling with this abundance and with this
overwhelming number of different apps just on the positive side there was a lot of worry
early on that texting would ruin young people’s language and make them poor writers and this
hasn’t been studies very systematically but the few studies we have show that young people
who text a lot actually have better language skills the play that they do with language
actually makes them better writers so I think there’s hope there I’m impressed with the
level of intellectual sophistication with which adolescents approach deciding where
to post what because they’re all moving across multiple platforms and they are very strategic
on what they put on Facebook where their parents or ministers or coaches might see what they
tweet which is a whole different audiences what they put on their personal blogs I think
that takes a lot of cognitive sophistication and I’d love someone to study what this is
all doing to their theory of mind because they have to constantly think about who’s
reading this who am I talking to that takes a lot of cognitive sophistication and I think
it may be building some positive skills all chime in as well building on Marion’s point
yes this is a very highly cognitively demanding set of things to navigate it’s certainly is
not the source of issues that our brain evolved to navigate so the interesting question for
me is in principal these sorts of technologies are democratic and egalitarian principal now
everybody has access to information everybody has access to elite college courses in actual
fact I think we’ve seen exacerbation of inequality due to this technology some people are increasingly
sophisticated at this sort of thing some people are less so and my I think most important
if I had to list the thing that I think is most important are we all going to be able
to leverage this sort of technology to help bring up the people who might have the least
privileged or is it just going to super charge the people who started off as privileged I
have to admit I’m less optimistic than I would like to be Andy than you for talks I guess
my question is on the topic of abundance but not among young people or tech users but about
researchers of relationships the internet gives us a fire hose of people’s most intimate
thoughts believes their drug purchasing strategies I’d like to know from each of you what you
think the greatest ethical or science integrity challenge you think will be in the next ten
years given that this is the direction that things are taking I think the greatest ethical
challenge we are going to have on this front is the movement toward data sharing I think
that there are large swaths of the field that really should not be posting this stuff at
all and I think that the discussion of like research ethics there seems to be like one
view of what ethical searcher is its more openness and right everybody can agree that
as a general principal openness and disclosure are good things but I actually feel like if
we’re not careful we’re going to set up norms and possibly even rules that run the risk
of sever privacy violations and I have found that the people who are most actively advocating
for these perspectives to my mind are insufficiently attentive it’s like yeah yeah yeah you can
just protect it but I mean I have seen the private parts of our major movie stars and
I’m not upset about that personally but I probably wouldn’t have see some of these things
like the major banks have had these sorts of failures the NSA has let out information
like I don’t think we can be naive at all about which data can be cracked so I think
by far the biggest ethical threat to the sorts of data collection that people like us study
is overzealous adherence to the movement toward openness I think user consent and figuring
out ways of insuring that people have some sense of what’s happening with their data
and some control over it I don’t know if that the informed consent as we currently are using
it I don’t know if it works when you hand a piece of paper to an undergraduate when
they take a survey and I’m pretty sure it doesn’t work when you ask them to click a
button online so I think we’ve seen a number of instances where users don’t understand
what you know the algorithm by which for instance Facebook is attempting to help them with the
abundance problem in sorting the news fees so I think that and I think this goes back
to Eli’s point as well about the importance of skills right so understanding that kind
of under the hood dynamics on the part of users participants and that they have some
sense of what’s happening with their data I think is something we haven’t even begun
to figure out I agree that the future is already now what I think is not about research my
husband and I last week we were both booking a hotel room at a certain point and we had
both our laptops and Then he had to pay two hundred dollar more for the same hotel room
than I had to pay and it was because there was a cookie on his computer so I think the
thing is that privacy is one thing but they combine when systems combine all the data
from different sources that’s one of the most riskily future issues let me just add a thought
here too and I think uncritical endorsement of these technologies is a huge issue and
I’ll give you just an example from personal experience I’ve been doing some consulting
for the US Army and the Army simply has established Skype outlets s that soldiers who are away
from home can communicate with the family back home this turned out to increase divorces
why did it increase divorces because here you have somebody who is out in the field
and they’re communicating with the person back home after the five minutes or so of
the oh I miss you so much it got in damn the roof is leaking and there’s no one here to
help me fix it and I can’t control the kids so will you please tell them to do their homework
and it actually in and the guys are off in the field and you know their worries about
being killed and I think there is a little too much uncritical acceptance you know you
can’t get a new drug through the FDA without doing careful randomized trials of this kind
of thing how come we don’t have randomized clinical trials of an you these things to
see what their impacts are we have just a few more minutes are there any other questions
okay the microphones coming well you touched on this a little bit already but I wanted
to ask you to expand this you’re getting a lot of data but people using social media
are also getting data about themselves and their relationships almost quantitative data
like the number of likes and also with the online matching you get quantitative data
about your potential relationships what do you think that changes introducing this quantitative
element I think it allows for more I don’t want to same gaming the system but I want
to say more conscious manipulation right so I’ve done some work looking at memorial pages
on Facebook and one of the things that we found was that people who created these pages
ended up then seeing the number of likes or the number of members in this particular group
as some kind of indication of how worthwhile the life of this departed person was and Alice
Marwick and Danah Boyd have also done a lot of great ethnographic work kind of thinking
about this quantification of self so I think that there is you know that it enables this
kind of feed back where buy we can you know put up a picture and if it doesn’t get the
requisite number of likes we can take it down and not have it threaten our self esteem on
the other hand there’s also some very nice work that Catalina Toma Wisconsin has done
where she’s looking at the way in which profile creating and being on one’s profile can actually
have this self affirming process whereby we it kind of allows us to feel better about
ourselves after engaging after seeing the response from our network so I guess I’m not
I think it depends would be kind of the copout answer to that I worry a little bit about
a very old idea that one that I think is borne out by the little data we have and that’s
the rich get richer the poor get poorer hypothesis that people who have strong relationships
and face-to-face interaction are also pretty social competent and adept in what they do
online and those who are less adept who may be more lonely or anxious or less social skilled
are also that way online in ways that invite further rejection I think we see some of that
in our work with adolescents I’m especially seen it in a project with thirteen-year-olds
I think there the quantification can be very harmful it could be really distressing to
those who keep putting themselves out there but don’t get the peer affirmation that they
see their more competent peers getting often when doing the very same things maybe just
with a little shade of difference yeah one more question so it’s another philosophical
one I guess so it’s about the children and the new generation so cumulative and humans
and have like no opportunity to control it and we just need to deal and research the
different affects and influences on our society but thinking about the next generation of
children do you think we will need psychologists need to think maybe how to provide specific
education on not just digital literacy but education about adoption of these technologies
and their usage in their social life like for example internet etiquette or which things
do we need to teach to our children to maybe protect them or so what’s a psychologist to
do given to that fact you know we just agreed that for children on and offline communication
have merged so I think education should of course should also include online environment
of kids and you know we talked about the main issues stayed the same but you know at an
early age you know you see that the things are rehearsed and practiced and explored at
an earlier age so what you now see before preadolescences when they start to get interested
in intimate topics we need to educate and we need to be part of general parenting I’d
think you’d agree on that oh absolutely AI think we do need to socialize children in
the digital domain I think a lot of parents are afraid and we give up in that domain and
it makes no sense we socialize children in every other context in which we send them
and I think we need to socialize them in their relationships online as well as offline and
I think we are out of time thank you all so much for coming and thanks to APS for organizing
this thank you so much

Leave a Reply

Your email address will not be published. Required fields are marked *