This study was conducted between September 1, 2009 and September 15, 2010 and received IRB approval from both the University of Wisconsin and the University of Washington.
Setting and subjects
This study used the SNS Facebook (www.Facebook.com) as it is the most popular SNS among the target population of college students.28, 29 Publicly available Facebook profiles of undergraduate students who were members of two large state university Facebook networks were investigated. Profile owners were selected for the study if their reported age on the profile was between 18 and 20 years old and the profile showed evidence of activity in the last 30 days. Only profiles for which profile owners could be recruited to the study by calling a phone number listed on either the Facebook profile or the university directories were included.
Eligible profiles were identified by a random search of the freshmen, sophomore, and junior undergraduate classes at the two selected universities using the Facebook search engine. Because the Facebook search engine is subject to error, all profiles returned in the searches were reviewed manually to confirm inclusion criteria were met. Profiles were excluded if they did not meet search criteria, including those who were not undergraduates (n = 448), did not meet the age criteria (n = 313), or did not display their age (n = 49). Profiles were also excluded due to privacy settings that prevented identification of depression symptoms, including having any one of the following sections set to private: information section, wall, or photographs (n = 1,630), or if a profile examination revealed that they would not be reachable for recruitment as no contact information (phone number or email) was listed on the profile or in the university directory (n = 303). A total of 307 profiles met all inclusion criteria and were evaluated.
Codebook and variables
All eligible profiles were evaluated by one of three trained coders using a SNS research codebook, which was developed and used in previous work evaluating displayed depression symptom references on SNS profiles.23, 25 From each SNS profile that met inclusion criteria, demographic data and displayed depression symptom reference data were recorded, including verbatim text from profiles. If present, identifiable information was removed from text references. References to depression symptoms were defined using the DSM-IV symptom criteria for a MDE.30 The criteria for MDE included depressed mood, loss of interest/pleasure in activities, appetite changes, sleep problems, psychomotor agitation or retardation, energy loss, feeling worthless or guilty, decreased concentration or suicidal ideation.30 Status updates were considered a depression symptom reference if they fit one of the described depression criteria by keyword or a synonym. For example, one symptom keyword of major depression is “hopeless,” therefore a status update stating “I feel hopeless” would be coded as a reference to depression. The term “giving up” is a synonym of “hopeless,” therefore, a status update disclosing “I feel like giving up” would be coded as a reference to depression. Status updates that clearly referenced a person other than the profile owner (i.e.“Matt is sitting next to me in class and he looks sad”), or references to the common situational experience of having a bad day (i.e. “I'm having a bummer of a day”) were not considered depression references.
Profiles were categorized into one of two groups. Profiles without any depression symptom references were considered “non-displayers.” Profiles with one or more references to depression symptoms were considered “Depression Symptom Displayers.”
First, the investigators viewed the “information section” of the profile to obtain demographic information and descriptive variables describing Facebook use. Second, to evaluate depression symptom disclosures, investigators reviewed each profile's status updates from the date of evaluation through the same date 1 year prior. For each status update that included a depression symptom, coders assessed whether the displayed symptom was a match to one of the DSM-IV criteria for a MDE.30 If so, they recorded verbatim text and the date of disclosure.
A 20% random subsample of profiles were evaluated by all three coders to test interrater reliability. Cohen's kappa statistic was used to evaluate the extent to which there was overall agreement in the coding of the presence or absence of depression symptom references on a profile. Cohen's kappa was 0.79 for depression symptom references.31
All profile owners whose profiles met inclusion criteria were telephoned. After verifying the profile owner's identity, the study was explained to the profile owner and permission was requested to send them an email that contained further information about the study. If the participant consented to receive the email, an email was sent to the profile owner's university email account that provided detailed information about the study as well as a link to the online survey. Two different links to identical surveys were used so as to obtain aggregate results without requiring personal identifiers. Subjects in the depression symptom displayer group were emailed a link to one survey; subjects in the non-displayer group were emailed a link to a separate but identical survey. The survey was administered online via a Catalyst WebQ online survey engine. Survey respondents were provided a $15 iTunes gift card as compensation. A total of 307 profiles met inclusion criteria and from these, 224 eligible profile owners participated in the survey (73% response rate) and 215 participants completed all PHQ-9 questions and received a total score.
The survey evaluated depression using the Patient Health Questionnaire (PHQ-9) clinical screen for depression. This screen is based on DSM-IV criteria for a MDE and has been validated in adult and adolescent populations.5, 32 The scale inquires about frequency of depression symptoms experienced in the last two weeks, such as depressed mood and hopelessness. Response categories include: not at all, several days, nearly half the days, and nearly every day. PHQ-9 scores range from 0 to 27; a score of less than 5 suggests no depression, a score of 5 or greater suggests depression. Depression diagnostic categories are differentiated as follows: a score of 5–10 suggests the person has mild depression, a score of 11–15 indicates the person has moderate depression, and a score over 15 suggests moderately severe depression.
All statistical analyses were conducted using STATA version 11.0 (Statacorp, College Station, TX). Displayed depression categories from Facebook data and demographic characteristics were summarized using descriptive statistics. Bivariate comparisons were performed using Fisher's exact test or Chi-squared tests. To examine associations between displayed depression references on Facebook and PHQ-9 score, two analyses were performed. First, zero-inflated negative binomial regression was used to examine the relationship between Depression Symptom Displayer/Non-Displayer as the predictor variable and PHQ-9 mean score as the outcome variable. Second, in an exploratory analysis, PHQ-9 scores were dichotomized to represent standard clinical categories: scores of 0 to 4 were considered not indicative of depression; any score 5 or above was indicative of depression (including mild, moderate, and moderately severe depression). Logistic regression was used to evaluate the relationship between depression symptom displayer/non-displayer as the predictor variable and the dichotomized PHQ-9 depression category as the outcome variable.