Abstract
Geosocial networking phone applications (GSN apps) used to meet sexual partners have become popular in the men who have sex with men (MSM) communities of the USA since 2009. Previous studies have shown greater incidence of gonorrhea and chlamydia and lower testing frequency for HIV testing among GSN app users when compared to non-users. The present study aims to determine the type, number, and frequency of apps used, as well as the association between dispositional self-control and health-related behaviors. Participants (n = 146) were recruited from Amazon’s Mechanical Turk program to respond to a brief GSN app marketing survey. Multivariate regression was used to determine differences in app priorities, length of app use, and number of sexual partners between high self-control and low self-control participants. A majority of the participants reported using Grindr (78 %) followed by Scruff (19 %), Growlr (12 %), and Jack’d (12 %). Most individuals used one app only (58 %), but a significant proportion reported using two apps (28 %) or three or more apps (14 %). Respondents with low self-control were more likely to report a higher number of hours using GSN apps and a higher number of sexual partners, controlling for race/ethnicity, education, employment, and HIV status. Given the popularity of this burgeoning communication medium, these findings have important implications for developing prevention resources for different segments of GSN app users.
Keywords: Self-control men who have sex with men, Social networking apps, Geosocial networking apps
Introduction
In March of 2009, a Los Angeles-based company named Grindr launched a new type of social network where new connections were not made through personal connections (e.g., Facebook) or professional connections (e.g., LinkedIn) but through sheer geographical proximity to other users. This geosocial networking app (GSN app) utilizes the smartphone’s internal global positioning system to map the user’s location in relation to other users, with those closest in proximity appearing first.
Catering to gay, bisexual, and other men who have sex with men (MSM), Grindr experienced a rapid uptake in a community where identifying potential sexual partners has historically been risky, especially in areas with high levels of homophobia.1 In 2013, the company reported that it had over four million users in 192 different countries.2 Other developers took notice of Grindr’s meteoric rise, and me-too apps like GROWLr, BoyAhoy, MR X, and SCRUFF were established to cater to different segments within MSM community. For example, GROWLr advertises in the Google Play store that users can meet “gay bears near you”3 with the term “bear” referring to a gay man that both is larger in size or weight and has an above average amount of body hair.
The proportion of MSM using these apps has increased substantially between 2011 and 2013. In August 2011, approximately 14 % of MSM clients receiving HIV and sexually transmitted disease (STD) screening services at the Los Angeles LGBT Center reported recently using GSN apps to meet sexual partners. In August 2013, the proportion of MSM receiving screening HIV/STD services who reported GSN app use to meet sexual partners increased to 24 %.4 Although Los Angeles is only one area in the USA, the increased user base for Grindr and the emergence of competing apps show that there is substantial demand in the MSM community for GSN apps.
While GSN apps may be increasingly popular, including also among heterosexual populations (e.g., Tinder), increased efficiency in meeting sexual partners could possibly lead to increased sexual risk-taking behavior and adversely affect the sexual health of MSM using these apps. A study by Rendina et al. found that a greater proportion of Grindr users had never been tested for HIV when compared to a population-based sample of MSM.5 Another study by Landovitz et al. in Los Angeles found that 46 % of MSM using Grindr reported unprotected intercourse in the past 3 months.6 A study by Lehmiller and Ioerger found that app users reported a significantly higher prevalence of ever being diagnosed with an STD compared to non-users.7 Last, a study by Beymer et al. showed that individuals using Grindr had a greater odds of contracting both gonorrhea (AOR = 1.25) and chlamydia (AOR = 1.37) when compared to individuals who solely used in-person methods like going to bars or clubs to meet sexual partners.8
It is unclear how the dispositional tendency to exercise self-control may influence sexual behavior when interfacing with these GSN apps. In 1992, Exner et al. found that MSM with low self-control had a greater number of sexual partners and were more likely to use drugs during sex.9 More recently, Adam et al. showed that sexual self-control was protective against unprotected anal intercourse with casual partners, but it was not protective on number of sexual partners or infections with an STD.10 While both of these studies present important findings, they were both conducted before the advent of smart phone apps and thus did not analyze domains unique to app use. Lehmiller and Ioerger found no differences between app users and non-app users in self-control. However their analysis did not look at within group differences.7 Thus, it is unclear how self-control impacts the number of apps used, the length of app use, or the number of sexual partners resulting from meetings facilitated by these apps. Furthermore, it is unclear if self-control scores may be associated with prioritization of viewing a potential user’s HIV/STD results and thus ability to navigate safely through this electronic medium.
The objectives of this study are threefold: (1) determine the type, number and frequency of GSN apps used among a sample of MSM in the USA (2) describe how trait-based self-control is related to app user behaviors, and (3) determine if there’s a link between self-control and prioritization of HIV/STD results. Elucidation of how these apps are used and the priorities for app users could reveal important information that can be used by app developers to maximize the safety of the user experience.
Materials and Methods
Data Collection
Participants were recruited from Amazon’s Mechanical Turk (MTurk) platform (Seattle, WA) from November 24th, 2013 to May 19th, 2014. MTurk is an online survey engine that allows researchers, called requesters, to post surveys on different topics from marketing to public health. Individuals sign up to become workers and are paid funds to complete surveys called Human Intelligence Tasks (HITs).11 These completed HITs can be either approved or rejected by the requester, and payment is conditional upon approval. Provided a worker’s HIT is approved, individuals are paid a credit to their Amazon.com account for each HIT taken. These credits can then be redeemed for products on the website or transferred to an individual’s bank account. In order to preview the HIT for this study, MTurk workers were required to be (1) located in the USA and (2) have a HIT approval rate of greater than or equal to 95 %. These measures were intended to increase data validity and maintain a homogenous sample of MSM within the USA.
Within the survey description, participants were instructed to complete the survey using a link posted on the MTurk HIT. The survey was programmed using Qualtrics software (Provo, UT), a platform that provides easy to use survey creation. The Qualtrics’ “prevent ballot box stuffing” option was used in order to prevent individuals from taking the survey multiple times.
Participants were instructed to complete a four question screening tool to determine their eligibility for the study. Participants were allowed to complete the full survey if they met the following inclusion criteria: (1) birth sex and current gender of male (subsequently referred to as cis-gender male), (2) sex with at least one man in the past 12 months, (3) currently owned a smart phone, and (4) had used GSN smart phone apps to socialize with other gay and bisexual men in the past month. A total of 652 participants were screened for this survey, and 245 met the above inclusion criteria for participation.
Participants not meeting the screening criteria were redirected to a disqualification page. The reasons for ineligibility were incompletion of screener (n = 56), non cis-gender male identification (n = 63), no sex with another man in the last 6 months (n = 140), did not own a smart-phone (n = 24), did not use GSN apps to meet other men in the past month (n = 119). All participants were paid $0.50 for their participation, regardless of passage of the screening criteria.
Although the core set of questions on GSN app behaviors and preferences remained the same, additional questions were added in a second phase of the study that asked about self-control, depression, compulsivity, and demographics (n = 74 in Phase I; n = 171 in Phase II). Therefore, a total of 245 individuals took the survey, but only 171 were asked the expanded questionnaire set. This study focuses only on the individuals who completed all questions.
After completing the screening assessment, participants were asked to list the specific GSN apps they used to socialize with other gay and bisexual men. Each response was checked for validity to ensure that the app(s) listed met the criteria for a GSN app. Twenty-five individuals were excluded during this inspection since the specific apps mentioned did not include GSN apps (e.g., Google, iTunes, WhatsApp) for a final sample size of 146.
The remainder of the survey asked a series of demographic and marketing questions about the number of apps used, the frequency of use, and willingness to pay for an upgraded version of these applications. Participants were then asked a series of rank-order questions which instructed them to rank their most preferred (1) app attributes, (2) profile attributes, and (3) convenience items, with each category containing five options where 1 indicated the highest priority and 5 the lowest priority.
These features were selected from the most common attributes and software features mentioned in 100 random reviews by individuals posting reviews in the iTunes and Android marketplaces for the following apps: Grindr, Scruff, Jack’d, and BoyAhoy. These features were separated into three categories. The first category, app attributes, is a general category that concerns the price, customer service, and software glitches. Profile attributes made up the second category and is tied to the frequency of different items including users, number of visible users, and number of pictures. The third category was titled “convenience items” which connotes attributes related to “ease-of-use” and users’ impatience.
The final section of the assessment asked about self-control, depression, compulsivity, and demographics. The Tangney Self-Control Scale was used to assess self-control.12 The original 36-item scale was shown to be reliable in this sample with an internal consistency of α = 0.89.
Statistical Analyses
A dichotomized self-control variable was created using a spotlight analysis approach to split the variable at the 25th and 75th percentiles. Briefly, spotlight analysis creates focal points (e.g., high self-control and low self-control) in order to ease interpretation of results.13 Following creation of this variable, independent samples t tests were used to assess the relationship between this dichotomized self-control measure and a set of five: (1) app attributes; (2) profile attributes; and (3) convenience items.
Following this analysis, the continuous measure of the self-control variable was used to test whether self-control was a significant predictor of the number of apps used, length of app use, number of sexual partners, prioritization of HIV/STD results, and likelihood of paying for additional features in multivariable regressions. Poisson regressions were used for count outcomes and logistic regressions were used for binary outcomes. For Poisson models, overdispersion was checked and goodness-of-fit Chi-square tests were used to check model fit. All models controlled for race/ethnicity, education, employment, and HIV status. All analyses were performed using SAS version 9.3 (Cary, NC).
Results
On average, it took 5 min and 24 s to complete the survey (interquartile range, 2:48 to 7:04) for individuals who answered all questions (n = 146). Participants who answered the demographics questions most frequently identified as White (75 %) followed by Hispanic (8 %), African-American (6 %), Other Race (8 %) and Asian/PI (4 %; Table 1). Over one third (38 %) reported a college degree for educational attainment, 36 % reported some college, 14 % reported graduate study, and 12 % had less than a college degree. A majority of participants reported full-time work (60 %) and part-time work (17 %), but a sizeable proportion was not employed (19 %). Approximately 55 % of participants reported their relationship status as single, 20 % of participants reported having a main partner and casual partners, and 25 % of participants reported having a main partner with no casual partners. Of those who responded to the HIV status question, 91 % were HIV negative, 5 % were HIV positive, and 3 % did not know their HIV status. Approximately 59 % of participants reported using only one GSN app, 29 % reported using two apps, and 12 % used three or more apps. Participants reported a wide geographic distribution with 36 of the 50 states represented in the USA.
TABLE 1.
Demographics of included participants (n = 146)
| Demographic | n | % |
|---|---|---|
| Race/ethnicity | ||
| White | 109 | 74.7 |
| African-American | 9 | 6.2 |
| Hispanic | 11 | 7.5 |
| Asian/PI | 6 | 4.1 |
| Other | 11 | 7.5 |
| Education level | ||
| High school graduate/GED | 17 | 11.6 |
| Some college | 53 | 36.3 |
| College degree | 55 | 37.7 |
| Post-graduate study/degree | 21 | 14.4 |
| Employment status | ||
| Employed full-time | 88 | 60.3 |
| Employed part-time | 25 | 17.1 |
| Not employed | 28 | 19.2 |
| Missing | 5 | 3.4 |
| Relationship status | ||
| Single | 80 | 54.8 |
| Main partner with casual partners | 29 | 19.9 |
| Main partner with no casual partners | 37 | 25.3 |
| HIV status | ||
| Positive | 7 | 4.8 |
| Negative | 133 | 91.1 |
| Don’t know | 4 | 2.7 |
| Missing | 2 | 1.4 |
| Number of geosocial network apps used | ||
| 1 app | 86 | 58.9 |
| 2 apps | 42 | 28.8 |
| 3 apps or more apps | 18 | 12.3 |
| State of residence | ||
| Alabama | 4 | 2.7 |
| Arizona | 2 | 1.4 |
| California | 19 | 13.0 |
| Colorado | 1 | 0.7 |
| DC | 2 | 1.4 |
| Florida | 9 | 6.2 |
| Georgia | 5 | 3.4 |
| Hawaii | 1 | 0.7 |
| Illinois | 12 | 8.2 |
| Indiana | 1 | 0.7 |
| Iowa | 1 | 0.7 |
| Kentucky | 6 | 4.1 |
| Louisiana | 2 | 1.4 |
| Maryland | 3 | 2.1 |
| Massachusetts | 2 | 1.4 |
| Michigan | 4 | 2.7 |
| Minnesota | 1 | 0.7 |
| Missouri | 3 | 2.1 |
| Montana | 1 | 0.7 |
| Nebraska | 3 | 2.1 |
| Nevada | 3 | 2.1 |
| New Hampshire | 1 | 0.7 |
| New Jersey | 7 | 4.8 |
| New York | 7 | 4.8 |
| North Carolina | 6 | 4.1 |
| Ohio | 6 | 4.1 |
| Oklahoma | 1 | 0.7 |
| Oregon | 2 | 1.4 |
| Pennsylvania | 9 | 6.2 |
| Rhode Island | 1 | 0.7 |
| South Carolina | 2 | 1.4 |
| Tennessee | 2 | 1.4 |
| Texas | 12 | 8.2 |
| Utah | 1 | 0.7 |
| Washington | 1 | 0.7 |
| Wisconsin | 1 | 0.7 |
| Unknown | 2 | 1.4 |
| Length of daily geosocial network app use | ||
| Mean | 42 min | |
| Standard deviation | 62 min | |
| Number of sexual partners in the past month | ||
| Mean | 2.86 sex partners | |
| Standard deviation | 4.18 sex partners | |
| Self-control outcome | ||
| Mean | 3.18 | |
| Standard deviation | 0.53 | |
| Total | 146 | 100.00 |
The mean amount of daily GSN app use was 42 min (SD = 62 min). The mean number of sex partners in the last month was 2.86 (SD = 4.18). The range of values for self-control was between 1 and 5 with participants averaging a score of 3.18 (median = 3.08; standard deviation = 0.53). The most common apps used were Grindr (78 %) followed by Scruff (19 %), Growlr (12 %), and Jack’d (12 %) (Table 2).
TABLE 2.
Type of geosocial networking apps used in the past month by participants (n = 146)
| App(s) used | n | % |
|---|---|---|
| Adam4Adam/Radar | 10 | 6.8 |
| BoyAhoy | 4 | 2.7 |
| Craigslist | 1 | 0.7 |
| Grindr | 114 | 78.1 |
| Growlr | 17 | 11.6 |
| GuySpy | 1 | 0.7 |
| Hornet | 11 | 7.5 |
| Jack’d | 18 | 12.3 |
| Manhunt | 2 | 1.4 |
| Mister | 2 | 1.4 |
| Recon | 2 | 1.4 |
| Scruff | 28 | 19.2 |
| Skout | 5 | 3.4 |
| Tinder | 7 | 4.8 |
| Othera | 8 | 5.5 |
| Total | 146 | 100.0 |
App use is not necessarily mutually exclusive since an individual can use more than one app. Therefore, the numbers exceed 146 when cumulatively added
aOther were apps that were only mentioned once or twice (Grommr, Squirt (2), Bender (2), GayGuysMeet, Gayvox, Krave, OKCupid (2), VGL, iDate, Plenty of Fish (2))
Independent samples, t tests revealed that prioritization of price of initial download (p = 0.62), log-on requests (p = 0.25), speed that the company fixes glitches (p = 0.64) and quality of customer service (p = 0.74) were not significantly different between individuals in the low self-control and high self-control groups (Table 3). Minimal loading time was significant at the 0.05 level (p = 0.05).
TABLE 3.
Spotlight analysis of mean scores for rank order questions and independent sample t tests on self-control and measured attributes (n = 80)
| Attribute | Low self-control | High self-control | p value |
|---|---|---|---|
| App attributes | |||
| Price of initial download | 2.05 | 1.90 | 0.62 |
| Minimal loading time/speed of use | 2.23 | 2.73 | 0.05 |
| Speed that the company fixes bugs/glitches in the software | 3.45 | 3.33 | 0.64 |
| Quality of customer service | 3.63 | 3.73 | 0.74 |
| Stays logged in/few log-on requests | 3.65 | 3.33 | 0.25 |
| Profile attributes | |||
| Number of men using the app | 1.7 | 1.73 | 0.93 |
| Number of other users you can see | 2.38 | 1.90 | 0.03 |
| Number of pictures you can post or view | 2.98 | 3.20 | 0.26 |
| Number of profiles you can add to your favorites | 3.9 | 4.33 | 0.06 |
| Number of other users you can block | 4.05 | 3.85 | 0.41 |
| Convenience items | |||
| Ease of uploading or sending pictures | 2.25 | 3.03 | 0.004 |
| Ease of filtering out non-local user profiles and messages | 2.95 | 2.33 | 0.04 |
| Ease of seeing another user’s HIV/STD results | 2.95 | 3.00 | 0.87 |
| Ease of filtering out offline profiles | 3.28 | 3.50 | 0.49 |
| Ease of blocking ads | 3.58 | 3.15 | 0.21 |
In regard to profile attributes, high self-control users reported placing significantly greater importance of the number of users one could see, when compared to low self-control users (p = 0.03). However, there were no significant mean differences detected between low self-control and high self-control users on the number of men using the app (p = 0.93), number of pictures to post/view (p = 0.26), number of blocks (p = 0.41), or number of profiles to add to favorites (p = 0.06).
In regard to convenience items, as expected, respondents with low self-control reported higher prioritization for uploading/sending photos (p = 0.004) but lower prioritization for filtering out non-local user profiles (p = 0.04). There was no mean difference between low self-control and high self-control respondents on seeing another users STD/HIV results (p = 0.87), ease of blocking ads (p = 0.21), or ease of filtering out offline profiles (p = 0.49). There may be a difference among these attributes, but possibilities for a type II error given the sample size may have prevented detection.
Multivariable Poisson regressions revealed that (positive) self-control was significantly inversely associated with the length of app use (p = 0.02) and number of sexual partners (p < 0.0001), controlling for race/ethnicity, education level, employment, and HIV status. However, multivariable logistic regression models showed that self-control was not significantly associated with likelihood of paying for additional features (p = 0.15) or HIV/STD results prioritization (p = 0.46), when controlling for the aforementioned covariates (Table 4).
TABLE 4.
Multivariate regression models of self-control and length of app use, number of apps used, number of sexual partners, prioritization of HIV/STD results and likelihood to pay for additional features (n = 146)
| Dependent variable | Estimate | Standard error | Point estimate (95 % CI) | p value |
|---|---|---|---|---|
| Poisson regression results | ||||
| Length of app use | −0.28 | 0.12 | 0.76 (0.59–0.96) | 0.02 |
| Number of apps used | −0.09 | 0.13 | 0.91 (0.7–1.19) | 0.49 |
| Number of sexual partnersa | −0.85 | 0.15 | 0.43 (0.36–0.58) | <0.0001 |
| Logistic regression results | ||||
| Prioritization of HIV/STD results | 0.25 | 0.34 | 1.29 (0.66–2.53) | 0.46 |
| Likelihood to pay for additional features | 0.44 | 0.31 | 1.56 (0.86–2.83) | 0.15 |
Controlling for race/ethnicity, education level, employment status, and HIV serostatus
aCorrected for overdispersion
Discussion
The present study found that individuals with low self-control were significantly more likely to report a greater number of sexual partners and higher daily GSN app use on average than individuals with high self-control. Furthermore, this study found that self-control was not associated with the prioritization of viewing another user’s HIV/STD results or the number of GSN apps used.
The number of sexual partners among MSM has been shown to be an important predictor of HIV infection in the male gay and bisexual communities of the USA.14–17 Given that 92 % of individuals who reported their HIV status were HIV-negative, participants with low self-control may be at a greater risk for HIV infection with an elevated number of partners. Furthermore, the low mean prioritization of HIV/STD results, regardless of self-control level, implies that GSN app users may not be concerned with HIV/STD results of prospective sexual partners, therefore putting them at further risk for contracting HIV.
This study is subject to important limitations. The cross-sectional study design did not allow us to draw conclusions about longitudinal GSN app use behaviors. Second, the study had two phases which added additional questions in the second phase such as demographics. Since demographic questions were not asked in both phases, demographic comparisons were not possible to assess comparability of populations. Third, other outcomes possibly related to self-control were not measured such as substance use, frequency of unprotected intercourse, and sexually transmitted diseases. Future studies investigating self-control among MSM respondents should look at all significant items shown a priori to be related to sexual health among MSM.
Fourth, although the recruitment platform was nationally-based in the USA, internet surveys have historically been biased in regards to the demographics of the participants recruited.18–21 Despite this limitation, Amazon’s Mechanical Turk has been shown to be more demographically representative of the US population when compared to standard internet sampling platforms.22,23 In addition, internet surveys are also subject to higher rates of drop-out and repeat participations,24 but participants in this study were not given their compensation code until the end of the study and the “prevent ballot box stuffing” option in Qualtrics ensured that participants did not take the survey multiple times. A requirement of a 95 % HIT approval rate was used which likely increased validity of results but may have biased the sample towards individuals who were more likely to complete the task. The respondents were generated via nonprobability sampling which has been shown to be less representative than respondents generated through probability sampling.25 Although participants were self-selected, the screening tool was used to include only a defined population for this study.
GSN apps provide an effective way to meet other sexual partners, but their primary use does not necessarily preclude their simultaneous use as health platforms. Account registration could include a simple survey that assesses key psychosocial indicators, such as the brief Tangney self-control scale, that have been shown to be linked with risk of HIV/STD risk, and provide periodic resources based on an individual’s profile and frequency of use. Alternatively, GSN apps could use the global positioning system function to show users the location of HIV/STD testing centers within close proximity as a reminder to test periodically. Last, companies could provide added features at no cost for users who agree to answer periodic health surveys. The data from these surveys can in turn be used by companies to maximize the sexual health knowledge of their users.
Like bathhouses and internet hook-up sites before them, GSN apps are unquestionably here to stay in the MSM communities of the USA. However, as competition in the GSN app marketplace increases, users may demand more from GSN apps including greater investments in features that help users maximize the safety of their sexual encounters. This is best stated through the feedback of one of the survey participants, “These app[s] are interesting and I think there’s definitely a market for them, I just wish there was a way to better vet the people using the app.”
Given the on-going impact of HIV in the gay and bisexual male communities, GSN app providers must employ strategies to offer their services in a way that simultaneously caters to the needs of the user base while providing each user resources specific to their individual risk profiles that protects the health of the overall sexual network. GSN apps can and should act as health resources in addition to their primary function so that they can ensure an enhanced user experience and foster greater well-being among their users.
Acknowledgment
The authors acknowledge support from the Morrison Family Center at the UCLA Anderson School of Management. MRB was supported by the UCLA Postdoctoral Fellowship Training Program in Global HIV Prevention Research (Currier and Gorbach, PIs]); T32MH080634.
Compliance with Ethical Standards
Competing Interests
None.
Study Approval
This study was approved by the University of California, Los Angeles South General Institutional Review Board #5 (IRB00004474; project no. 13-001391).
Contributor Statement
MRB: analysis, critically revising manuscript and approval. AD: conception, design, analysis and interpretation of data, and critically revising manuscript. SBS: conception.
References
- 1.Signorile, M. Joel Simkhai, Grindr Founder and CEO, on gay social networking app’s worldwide impact. The Huffington Post. November 26, 2013. http://www.huffingtonpost.com/2013/11/26/grindr-joel-simkhai-_n_4343568.html. Accessed January 13 2014.
- 2.Grindr Website, “Learn more” http://grindr.com/learn-more. Accessed December 27, 2013.
- 3.Google Play, “GROWLr” https://play.google.com/store/apps/details?id=com.initechapps.growlr&hl=en. Accessed Sept 18 2015.
- 4.Los Angeles LGBT Center, Unpublished Data.
- 5.Rendina HJ, Jimenez RH, Grov C, Ventuneac A, Parsons JT. Patterns of lifetime and recent HIV testing among men who have sex with men in New York City who use Grindr. AIDS Behav. 2014;18:41–9. doi: 10.1007/s10461-013-0573-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Landovitz RJ, Tseng C, Weissman M, et al. Epidemiology, sexual risk behavior, and HIV prevention practices of men who have sex with men using Grindr in Los Angeles, California. J Urban Health. 2013;90:729–39. doi: 10.1007/s11524-012-9766-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lehmiller JJ, Ioerger M. Social networking smartphone applications and sexual health outcomes among men who have sex with men. Plos One. 2014;9 doi: 10.1371/journal.pone.0086603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Beymer MR, Weiss RE, Bolan RK, et al. Sex on demand: geosocial networking phone apps and risk of sexually transmitted infections among a cross-sectional sample of men who have sex with men in Los Angeles County. Sex Transm Inf. 2014;90:567–572. doi: 10.1136/sextrans-2013-051494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Exner TM, Meyerbahlburg HFL, Ehrhardt AA. Sexual self-control as a mediator of high-risk sexual-behavior in a New-York-City cohort of HIV-positive and HIV-negative gay men. J Sex Res. 1992;29:389–406. doi: 10.1080/00224499209551655. [DOI] [Google Scholar]
- 10.Adam PCG, Teva I, de Wit JBF. Balancing risk and pleasure: sexual self-control as a moderator of the influence of sexual desires on sexual risk-taking in men who have sex with men. Sex Transm Inf. 2008;84:463–7. doi: 10.1136/sti.2008.031351. [DOI] [PubMed] [Google Scholar]
- 11.Amazon.com, “Mechanical turk concepts” http://docs.aws.amazon.com/AWSMechTurk/latest/RequesterUI/mechanical-turk-concepts.html. Accessed August 10, 2015.
- 12.Tangney JP, Baumeister RF, Boone AL. High self-control predicts good adjustment, less pathology, better grades, and interpersonal success. J Pers. 2004;72:271–324. doi: 10.1111/j.0022-3506.2004.00263.x. [DOI] [PubMed] [Google Scholar]
- 13.Spiller SA, Fitzsimons GJ, Lynch JG, McClelland GH. Spotlights, floodlights, and the magic number zero: simple effects tests in moderated regression. J Mark Res. 2013;50:277–288. doi: 10.1509/jmr.12.0420. [DOI] [Google Scholar]
- 14.Buchbinder SP, Vittinghoff E, Heagerty PJ, et al. Sexual risk, nitrite inhalant use, and lack of circumcision associated with HIV seroconversion in men who have sex with men in the United States. JAIDS. 2005;39:82–9. doi: 10.1097/01.qai.0000134740.41585.f4. [DOI] [PubMed] [Google Scholar]
- 15.Brewer DD, Golden MR, Handsfield HH. Unsafe sexual behavior and correlates of risk in a probability sample of men who have sex with men in the era of highly active antiretroviral therapy. Sex Transm Dis. 2006;33:250–5. doi: 10.1097/01.olq.0000204746.58160.c1. [DOI] [PubMed] [Google Scholar]
- 16.Koblin BA, Husnik MJ, Colfax G, et al. Risk factors for HIV infection among men who have sex with men. AIDS. 2006;20:731–9. doi: 10.1097/01.aids.0000216374.61442.55. [DOI] [PubMed] [Google Scholar]
- 17.Schwarcz S, Scheer S, McFarland W, et al. Prevalence of HIV infection and predictors of high-transmission sexual risk behaviors among men who have sex with men. Am J Public Health. 2007;97:1067–75. doi: 10.2105/AJPH.2005.072249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ross MW, Tikkanen R, Mansson SA. Differences between internet samples and conventional samples of men who have sex with men: implications for research and HIV interventions. Soc Sci Med. 2000;51:749–58. doi: 10.1016/S0277-9536(99)00493-1. [DOI] [PubMed] [Google Scholar]
- 19.Rhodes SD, DiClemente RJ, Cecil H, Hergenrather KC, Yee LJ. Risk among men who have sex with men in the United States: a comparison of an internet sample and a conventional outreach sample. AIDS Educ Prev. 2002;14:41–50. doi: 10.1521/aeap.14.1.41.24334. [DOI] [PubMed] [Google Scholar]
- 20.Rhodes SD, Bowie DA, Hergenrather KC. Collecting behavioural data using the worldwide web: considerations for researchers. J Epidemiol Commun H. 2003;57:68–73. doi: 10.1136/jech.57.1.68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Evans AR, Wiggins RD, Mercer CH, Bolding GJ, Elford J. Men who have sex with men in Great Britain: comparison of a self-selected internet sample with a national probability sample. Sex Transm Inf. 2007;83:200–5. doi: 10.1136/sti.2006.023283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Buhrmester M, Kwang T, Gosling SD. Amazon’s Mechanical Turk: a new source of inexpensive, yet high-quality, data? Perspect Psychol Sci. 2011;6:3–5. doi: 10.1177/1745691610393980. [DOI] [PubMed] [Google Scholar]
- 23.Casler K, Bickel L, Hackett E. Separate but equal? A comparison of participants and data gathered via Amazon’s MTurk, social media, and face-to-face behavioral testing. Comput Hum Behav. 2013;29:2156–60. doi: 10.1016/j.chb.2013.05.009. [DOI] [Google Scholar]
- 24.Birnbaum MH. Human research and data collection via the internet. Annu Rev Psychol. 2004;55:803–32. doi: 10.1146/annurev.psych.55.090902.141601. [DOI] [PubMed] [Google Scholar]
- 25.Chang L, Krosnick JA. National surveys via RDD telephone interviewing versus the internet. Publ Opin Quart. 2009;73:641–78. doi: 10.1093/poq/nfp075. [DOI] [Google Scholar]