00:00:04
I conduct research related to
00:00:06
traditional and new media and in this
00:00:09
study what we are interested in is
00:00:10
documenting any sort of bias that might
00:00:13
be happening in newspapers related to
00:00:15
the availability of digital data and big
00:00:18
data consumers often wonder what goes
00:00:22
through an editor's mind when they look
00:00:24
at newspapers how do the news that they
00:00:27
read how do they make it to the
00:00:29
newspaper or the TV that they watch and
00:00:31
they often wonder are they selected
00:00:34
based on some sort of particular agenda
00:00:37
of the newspaper already sorted based on
00:00:39
their importance now if you ask a
00:00:42
journalist up until recently at least
00:00:44
they would give you the answer that yes
00:00:46
news are selected based on their news
00:00:48
worthiness there's their importance and
00:00:51
the novelty how important they are how
00:00:54
new they are to the public however if
00:00:58
this is the case of course consumers may
00:01:00
also wonder why is it that we see so
00:01:02
many cat pictures or hawise that we see
00:01:05
so many quizzes that are available on
00:01:07
the news media now we took an alternate
00:01:10
approach to this we had an alternate
00:01:11
hypothesis and we said but perhaps there
00:01:14
are other concerns in editors mind
00:01:16
especially these days when the editors
00:01:18
are fed so much information about how
00:01:21
the individual stories on online news
00:01:23
articles are doing in particular we are
00:01:26
thinking that there might be a
00:01:27
relationship between in the traffic that
00:01:29
individual articles receive and how long
00:01:32
stories particular topics are covered on
00:01:35
a newspaper so we tested this hypothesis
00:01:38
using data from a large online version
00:01:43
of a reputable newspaper outside of the
00:01:45
United States was actually an Indian
00:01:48
newspaper and we wanted to see if the
00:01:51
stories who's first articles received a
00:01:54
high number of clicks for various
00:01:57
reasons if they somehow were covered for
00:02:00
a longer period of time in the newspaper
00:02:02
and we also tested various additional
00:02:06
measures of resources that an editor
00:02:08
might allocate to the story and what we
00:02:11
found was interesting we indeed found
00:02:13
that there is a relationship between the
00:02:16
number of clicks that are
00:02:18
saved by the article and the amount of
00:02:21
time that a story is essentially covered
00:02:24
by the follow-up articles that are
00:02:26
repeating on multiple days those two are
00:02:30
positively correlated
00:02:35
well there are a couple key takeaways
00:02:37
from our research first is of course
00:02:39
this relationship that indeed whenever a
00:02:42
news article receives higher number of
00:02:44
clicks there is a higher level of
00:02:47
resources in tutorial resources that are
00:02:49
allocated to the particular story but
00:02:52
this is of course something that makes
00:02:54
sense at a correlational level you would
00:02:56
wonder to what extent is this also
00:02:59
causal the relationship what I mean by
00:03:02
that is to what extent is it the case
00:03:05
that just because new story is receiving
00:03:07
higher number of clicks you see a longer
00:03:11
period of time that's allocated to it so
00:03:13
we had to establish this this was the
00:03:15
the main effort that we exert exerted in
00:03:18
our research what we did was to find
00:03:22
some instruments as we call them that
00:03:25
would affect essentially the amount of
00:03:27
clicks that that are received by the
00:03:30
article but not necessarily the
00:03:32
editorial decision of how much time to
00:03:35
allocate to it and the two instruments
00:03:37
that we used were looking at reigns days
00:03:40
with reigns and looking also at the
00:03:43
ordinary power shortages so our idea is
00:03:46
that if there is a rainy day people
00:03:49
would be staying indoors they would be
00:03:51
more likely to go online that is going
00:03:53
to create some sort of a shock in the
00:03:55
number of clicks similarly on the days
00:03:57
when there are ordinary power shortages
00:04:00
it's likely that the chances that people
00:04:03
will go online is going to go down
00:04:05
that's going to reduce the number of
00:04:07
clicks in environment look at the
00:04:09
relationship between clicks and these
00:04:10
two instruments we find a significant
00:04:13
relationship and whenever we looked at
00:04:16
essentially the editorial policy related
00:04:18
to these two instruments we are also
00:04:20
seeing that indeed there is a
00:04:23
relationship that we can see in terms of
00:04:27
how long the coverage of a particular
00:04:29
stories or also on how many follow-up
00:04:32
articles appear related to a particular
00:04:34
first story that was published what
00:04:37
would that mean that might mean that if
00:04:39
it happens to be a rainy day or if it
00:04:42
happens to beat it in India if there was
00:04:44
an ordinary power shortage on the day
00:04:47
that
00:04:48
Kim Kardashian article first broke that
00:04:51
story actually may run for a couple more
00:04:54
days on the newspaper as opposed to
00:04:56
another story
00:05:02
well the conclusion that surprised us
00:05:04
the most was probably who once we
00:05:06
started separating news into further
00:05:08
classifications and and made a
00:05:10
comparison among them so what we did as
00:05:12
an addition was to analyze news as hard
00:05:15
versus soft news what I mean by hard
00:05:17
news is the type of news that tend to
00:05:20
provide more information to the
00:05:22
consumers for example news related to
00:05:24
politics or business and by soft means
00:05:27
soft news I mean news such as
00:05:29
entertainment or sports when you looked
00:05:32
at video traffic that's received by
00:05:35
articles influence the coverage of
00:05:37
stories by looking at only the hard news
00:05:40
in the soft news we found that the
00:05:42
effect is present only for the hard news
00:05:44
it doesn't exist for the softness or to
00:05:46
put it in more academic terms the effect
00:05:49
is not significant so what this might
00:05:51
imply is that consumers should not
00:05:53
necessarily worry that just because now
00:05:56
traffic information is provided to news
00:05:58
editors and news editors tend to pay
00:06:01
more attention to them it doesn't
00:06:03
necessarily mean they'll start seeing
00:06:06
newspapers full of cat pictures or more
00:06:09
entertainment stories in addition we
00:06:12
also found sort of we quantified the
00:06:14
effects of the traffic so what we can
00:06:18
say is for example increasing the new
00:06:22
story clicks by about seven hundred
00:06:24
klicks is going to result in about three
00:06:27
additional days of coverage in the
00:06:30
storage and that might also mean 3
00:06:32
additional articles on average about the
00:06:35
same story about the same topic so
00:06:37
that's the second fine
00:06:43
so in terms of the news industry right
00:06:47
now there's a lack of consensus in terms
00:06:49
of what the newspapers should be doing
00:06:51
and with the spit all of this digital
00:06:53
data when we look at some newspapers
00:06:55
walks calm is mentioned to be an example
00:06:58
of this they tend to recommend their
00:07:01
journalists to stay away from digital
00:07:03
beta this is done in order to encourage
00:07:06
them to come up with better stories
00:07:08
newsworthy stories stories that matter
00:07:11
for the society by me look at other
00:07:13
examples gawker.com is mentioned to be
00:07:16
an example of this these newspapers tend
00:07:20
to provide some sort of an incentive for
00:07:22
their journalists to create stories that
00:07:24
can bring higher number of clicks
00:07:26
usually these incentives are in the form
00:07:28
of financial payments so there isn't
00:07:31
really a consensus and it's not clear
00:07:34
exactly what policy should be the right
00:07:37
policy to embrace by the news industry
00:07:40
or study I think shows what can happen
00:07:43
or at least currently documenting what's
00:07:46
happening in one large reputable portion
00:07:50
of the news industry what we are telling
00:07:52
the news editors is in some sense yes it
00:07:56
might help to use the digital data as
00:07:58
long as you're not using this in order
00:08:01
to dilute the quality of information
00:08:02
that the consumers are receiving and for
00:08:06
the consumers I think we are telling
00:08:08
them that they don't have to necessarily
00:08:10
worry about the quality of information
00:08:12
that they receive from newspapers
00:08:14
digital data might still actually help
00:08:17
them it can increase the number of
00:08:19
stories that are informative in some
00:08:22
sense or the coverage of the stories
00:08:23
that are informative for policymakers
00:08:26
this brings up other questions because
00:08:29
the one other side of essentially
00:08:32
driving at clicks driving your news is
00:08:35
what happens to the representations of
00:08:38
the minority views when the majority of
00:08:40
clicks are received by the populations
00:08:43
or the sub groups that are also in in
00:08:46
larger portions in a society that might
00:08:48
bring up some questions about we point
00:08:51
diversity which is important from the
00:08:54
perspective of FCC
00:08:56
and they might need to start now asking
00:08:58
some questions about what does this
00:09:01
policy due to the viewpoint of some of
00:09:04
the minority maybe groups in the society
00:09:12
well technically our research is the
00:09:15
first that's documenting the existence
00:09:17
of bias that's coming from the traffic
00:09:20
so in that sense we can't really make a
00:09:21
comparison to form a research but we can
00:09:24
say that we are providing in evidence of
00:09:26
a bias that's in a different format and
00:09:29
that is essentially their new just
00:09:33
something that we are starting to see in
00:09:35
newspapers as an additional form of
00:09:38
traffic source twice
00:09:44
while we are still conducting and
00:09:46
carrying out further analysis to see if
00:09:49
we can make additional statements about
00:09:51
this type of research and we are getting
00:09:53
lots of feedback from or from our
00:09:56
scholars and also peers in the industry
00:09:58
as far as following up with our research
00:10:01
with the same research kotor i'm going
00:10:04
to be following now on social media not
00:10:07
just the traditional news media and we
00:10:09
are going to look at other impacts other
00:10:12
factors that might be influencing
00:10:14
people's life now one particular study
00:10:17
we are focusing on that's relevant but a
00:10:20
very different topic is looking at
00:10:22
politicians and three on social media
00:10:24
and how that influences their chances of
00:10:27
raising campaign donations
00:10:45
you
