Tuesday, December 15, 2009

Sexting: New Info about an Info Behavior

Pew Internet and American Life Project came out with a new report on "sexting" today. The basic findings: prevalence of sexting "ever" among teens overall is in the 10-20% range. Sexting seems to be an evolving element in teen "courtship behavior."

I was disappointed, though, with the "just-this-side-of-moral-crusading" feel of the report. The tone is not explicitly alarmist, but it is a soft ball pitch to those who will turn it into media hoo-ha.  Expect a number of misleading articles to appear in the media to be followed by researchers decrying media distortion.  But whose fault: consider the flaws in just this one report in terms of what we give the media to work with.

A Hesitance to Criticize Previous Research
As background they describe previous surveys, done by The National Campaign to Prevent Teen and Unplanned Pregnancy and National Center for Missing and Exploited Children and others. One found ~20% of teen participants had sent and ~30% had recieved a sexually suggestive picture or video of themselves to someone via email, cell phone or by another mode. In another 9% had sent, 3% had forwarded one, and 17% had received. All of these surveys seemed to have some methodological problems that would put wide-error bars on these numbers but the report just hints at these.

Slightly Fuzzy Numbers
This report is based on a survey of 800 young people plus focus groups.In the new study, the acknowledged margin of error for the full sample of 800 is about +/- 4%. For subgroups, it will be higher -- for the 1/6 sample of each age year, for example, it's about +/- 8%.

And then the report says
4% of all cell-owning teens ages 12-17 report sending a sexually suggestive nude or nearly-nude photo or video of themselves.... [among t]he oldest teens in our sample – those aged 17 – ... 8% ... hav[e] sent one, compared to 4% of those age 12.....
But given the margin of error, all we can say is that somewhere between 0 and 8% of all teens and somewhere between 0 and 16% of 17 year olds have sent a suggestive picture of themselves.  The authors do a great job of including background on the survey and footnoting margins of error and such but they leave it up to the savvy reader to make something of these.  All these numbers are pretty small -- this reader, at least, thinks responsible researchers should do a little more to drive home this point than this report does.

A Missing "Network" Angle
The authors don't make much of the fact that the number of folks who have sent is consistently lower than the number who have received. This implies, and their qualitative data seems not to deny, that the practice is not informally controlled by a norm of "just between you and me babe" and that the ease of distribution and the difficulty of detection and potential for sheer high volume make the transaction costs of informal control prohibitive.  Obvious, but important.

Percentaging in the Wrong Direction
The media pitch is furthered by doing percentages in arguably the wrong way. Consider this paragraph:
Teens who receive sexually suggestive images on their cell phones are more likely to say that they use the phone to entertain themselves when bored; 80% of sexting recipients say they use their phones to combat boredom, while 67% of teens who have not received suggestive images on their phone say the same. Teens who have received these images are also less likely to say that they turn off their phones when it is not otherwise required – 68% of receiving teens say they generally do not turn off their phones when they do not have to, and 46% of teens who have not received suggestive images by text report the same “always on” behavior (page 6).
As is, it risks being parody: those who receive naughty pictures are more likely to use their phones to combat boredom than those who do not! But presumably the point here is to compare types of cell phone users and so the percentages should be done the other way round: among boredom combatters, what percent get baudy pictures? A quick, back of the envelope recalculation* suggests it would look like this:









Use vs.
Boredom
Not vs.
Bordedom
Received ~108
(20%)
~27
(11%)

No Received ~445
(80%)
~221
(89%)


====
~553
====
~247




That's actually a little more compelling (and certainly easier to make sense of). The rate is twice as high among the "I use my phone to combat boredom" group. But both are relatively low.

A similar methods 101 error is made when reporting what interventions make sense:
One parental intervention that may relate to a lower likelihood of sending of sexually suggestive images was parental restriction of text messaging. Teens who sent sexually suggestive nude or nearly nude images were less likely to have parents who reported limiting the number of texts or other messages the teen could send. Just 9% of teens who sent sexy images by text had parents who restricted the number of texts or other messages they could send; 28% of teens who didn’t send these texts had parents who limited their child’s texting (page 12).






It is unlikely that the authors are thinking that sexting causes parental restrictions -- the sense is just the opposite -- and so the percentaging should be within the categories of parental behavior and comparison across these.  This should look like this (again, based on quick, back of the envelope, calculations*).






Parental
Restriction
No Parental
Restriction

Ever Sent ~3
(1.4%)
~29
(5%)

Never Sent
~215
(98.6%)
~553
(95%)


====
~218
====
~572




Again, this doesn't overturn the take-away -- it might even be argued that it strengthens it: lack of parental cell phone restriction associated with a 3 to 4 fold increase in the behavior -- but we researchers should put our best practices forward to as we dump our results and findings into the information environment around us.

* Calculations
Text says 
Just 9% of teens who sent sexy images by text had parents who restricted the number of texts or other messages they could send; 28% of teens who didn’t send these texts had parents who limited their child’s texting.
And
4% of all cell-owning teens ages 12-17 report sending a sexually suggestive nude or nearly-nude photo or video of themselves to someone else.
Thus: 0.04 x 800 = 32 were senders and 768 non-senders
Of senders, 0.09 x 32 had restrictive parents: = 2.88 = ~3
Of non-senders, 0.28 x 768 had restrictive parents: = 215
Total restrictive parents = 218, total non-restrictive 572


Text says
80% of sexting recipients say they use their phones to combat boredom, while 67% of teens who have not received suggestive images on their phone say the same.
Total sexters is 4% or 32.  80% of these is ~27 sexters who combat boredom.
Non-sexters are 768 and 67% of these is ~515.
Total boredom combatters is ~553, non-boredom-combatters is ~247.


Revisions
13 Feb 2012: fixed several typos and formatting glitches

2 comments:

  1. Nice summary! Unfortunately the Pew boys (and yes, they are almost uniformly boys), are pretty close to moral crusaders. And, they have a lot of company among the current crops of "survey researchers" with ties to right wing foundations and conservative Christian groups. Tenured sociologists at leading universities are now some of the key advocates for abstinence education, early marriage, and key opponants of sex between consenting adults (even heterosexual sex!). It's an embarrassment.

    ReplyDelete
  2. Anonymous6:49 AM

    I am reminded of a Rand survey I recently reviewed for my statistics class that also seemed to do this backward calculation: they surveyed the television viewing habits of teenagers and then attempted to correspond that to their pregnancy rates. However, the data they presented in the survey only included responses from sexually active teenagers, not the half of participants who had NO sexual activity. Basically, it came down to the exciting conclusion that, of teenagers who have sex, those who happen to watch a lot of the selected tv programs (no names were listed) were slightly more likely to get pregnant/get someone else pregnant. (As in the 90th percentile was twice the rate of the 10th percentile. Shocking!) But the media ran with it as tv making our kids get pregnant anyway.

    ReplyDelete