on plastic surgery requests

Cosmetic Design is the most frequently read cosmetics news website. They posted this article yesterday that sites the following:

Requests for surgery as a result of social media photo sharing rose 31 percent in 2012, reports the American Academy of Facial Plastic and Reconstructive Surgery.

Just a second thought, for this stat to actually hold water, they should have regressed people’s plastic surgery requests on their reasons for the requests. Instead the study Cosmetic Design was citing did not do this. The 2012 A.A.F.P.R.S Membership Study (download report here) asked 63 plastic surgeons in a survey between December 2012-January 2013 the following question:

Have you seen an increase in requests for plastic surgery stemming from people being more self aware of their looks because of social media? If so, estimate the percentage of requests for each procedure.

It was an opinion based question asked to plastic surgeons. Not a valid way to measure “rise.”

Many issues here: small sample, asking someone to estimate percentage increases seems a bit flawed (to me, at least), and asking the surgeon and not the client.

The 31% increase in plastic surgery requests due to social media photo-sharing is not the most reliable statistic.

This is what I would have done if I could re-design their study:

Data collection:

(1) Collect data on a similar sample of people before 2012. For example in 2008, collect data on women aged 20-45 in a given geographical region. Collect data on a similar sample of women in 2009, 2010, 2011, and 2012.

Questionnaire Design:

(2) The questionnaire should ask this sample if they are active users on a popular social media website.

(3) The questionnaire should also ask the following two questions: (1) if the individual is considering undergoing plastic surgery for cosmetics reasons this year and (2) if the individual has placed a request for plastic surgery for cosmetics reasons this year.

(4) Then the questionnaire should ask why the individual is considering undergoing plastic surgery for cosmetic reasons. One way to do this is have them answer yes or no questions to a list of reasons. One of these reasons should be:  “I want to look better in photographs” with the responses “Yes” and “No.”

Analytic Methodology:

For them to actually claim there is rise in plastic surgery requests because of social media photo sharing, this is how their analytic methodology should theoretically be:

(1) Take a sub-sample of women who have requested/considering plastic surgery for cosmetic reasons. Combine the variable and recode so there are only two responses: “Yes I have requested/considered plastic surgery” and “No, I have not request/considered plastic surgery.”  Doing this allows them to run a logistic regression. This is your dependent variable.

(2) Control for all the typical factors, age, education, martial status being the three that come to mind.

(3) Control for social media (if they have an account and are active) and control for all the reasons the individual is considering social media. These are your main effects.

(4) See if social media and “I want to look better in photographs” have significant betas. If they do, you can “kind of” support the because claim. Do this again for 2009, 2010, 2011, and 2012 data and then you kind of  support or dispute the rise claim.

(5) If any of the main effects are significant you can run an interaction. For example (a) say having a social media account and wanting to look better in photographs are both significant predictor of the dependent variable; (b) you can interact these two variables and see if women who have social media accounts and who want to look better in photographs significantly predicts the dependent variable. If it is significant then you have a juicy story.

Your turn : )

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s