Information about every aspect of personal and professional life is now available on the Internet; this is of course true for information about genetics and genomics. Collectively, genetic health professionals and genomic researchers use thousands of sites as reference points for information as well as spaces for discussion about issues pertinent to their industry, just to name a few: Decipher (www.decipher.sanger.ac.uk), GeneReviews (www.ncbi.nlm.nih.gov/books/NBK1116/) and GenomesUnzipped (www.genomesunzipped.org). People affected by genetic disease are able to easily access websites related to a condition or group of conditions, e.g. Unique (www.rarechromo.co.uk/html/home.asp) and GeneticAlliance (www.geneticalliance.org.uk). On Google the search term ‘genomics’ revealed 11 million hits; thus, the Internet offers an abundant data source and by virtue of this provides a rich viewing audience, ripe for collection for research assessing attitudes towards the use of genomics.
As interest in the survey could spread virally, i.e. people who enjoyed participating told their online friends about it, the sampling frame was thus both convenience and snowball. Due to this it was not possible to draw conclusions on whether the final sample was representative of the Internet population as a whole or indeed representative of any specific population.
The final ascertained sample consisted of participants who were predominantly female, white, highly educated and aged 31–50. Below is an exploration of whether this is a typical profile of people who take part in surveys as well as those who use social media, access traditional media such as news programmes and are part of the select professional groups targeted.
Demographics of social networkers
It is very difficult to obtain accurate information on the generic profile of Facebook, Twitter and LinkedIn users as the rate of growth for these three media is phenomenal and each site rarely reports user demographic data. It is also surprisingly difficult to mine the Internet generally for up-to-date statistics about social media that are evidence based, collected via robust research methods; thus, the following information is provided only as a guide.
The most popular age range for social media users generally is 35–44 years (Macmillan 2011); 65 % of US Facebook users and 37 % of UK Facebook users are 35 or older (Pingdom 2012). According to Sakki (2013) Facebook users are more likely to be over 25 (Sakki 2013). The average Facebook user is thought to range from 18–29 years (Duggan and Brenner 2013), 25–34 years (Fanalyzer 2013), 38 years (Macmillan 2011) through to 40.5 years old (Pingdom 2012). For Twitter, 55 % of US users are 35 or older (Pingdom 2012), and most Twitter users in the UK are over 35; the age range is between 18 and 29 years (Sakki 2013), and average age is 37.3 years old (Pingdom 2012) and 39 years old (Macmillan 2011). For LinkedIn, 79 % of US users are 35 or older, and the majority of UK users are over 35 (Sakki 2013) with the average user being 44.2 years old (Macmillan 2011; Pingdom 2012).
As Table 4 shows the 4,048 participants we recruited via social media were more likely to be in the 31–50 age range. Thus, our sample is typical of the ‘average’ user of social media as reported by other sources.
Women are more likely to access social network sites compared to men (Emerson 2011; eMarketer 2013), and according to the UK’s Office of Communications (Ofcom) those women who do access social media sites do so more frequently than men (Ofcom 2013). Women also have 55 % more wall posts on Facebook than men (Boglioli 2011), and women spend, on average, 9 % more in terms of time on social networking sites generally than men (Widrich 2013).
In the US 60 % of Facebook users are women (Pingdom 2012). In the UK 51 % of Facebook users are women (Fanalyzer 2013). In the US 60 % of Twitter users are women (Pingdom 2012), and for LinkedIn, 53 % are women (Pingdom 2012). Slightly different figures are given by Sakki in the UK, who report that Facebook is used in equal numbers by men and women, Twitter is used slightly more by women (51 % compared to 49 %) and LinkedIn is used more by men (58 % compared to 42 %) (Sakki 2013).
Table 4 shows that our sample recruited through social media was predominantly female. This also fits with the generic profile data on social media use by gender as reported by other sources.
Household income, education, ethnicity and marital status
The Pew Internet and American Life Project catalogues trends in social media use (www.pewinternet.org); this research relates to the American market and was taken from their latest survey in 2012. The average Facebook user is educated (73 % had some college attainment, and 68 % had completed college), with a household income above $75 k and living in urban areas (there was no data on ethnicity for Facebook; however, social media users generally were slightly more likely to be Hispanic or Black than White). Whereas the average Twitter user is African-American with some college education, with a household income above $75 k living in urban areas (Duggan and Brenner 2013). In the UK 69 % of Facebook users are in a relationship (Fanalyzer 2013). The majority of our sample recruited through social media were also in a relationship. Our sample was also overwhelmingly white (92 %), and there was little representation from other ethnic or racial groups. The vast majority of participants in the final sample were from Europe, and whilst this continent still consists of an eclectic mix of different ethnic and racial groups, the majority of people from Europe would still class themselves as white. We did not gather data on household income, but the profile of our users was of a very high level of academic achievement (70 % had a degree or higher level of education). Even if the health professionals and genomic researchers were removed from this calculation the research participants who are members of the public still selectively have a higher educational level than one might expect of a representative public.
Whilst generically it appears that social media users may be more likely to have higher education levels than not, our sample was particularly biased towards the well educated. This may be due to a combination of factors—the subject matter may hold particular interest to those who have studied biology before or to those who are interested in ethical issues raised by technologies. In addition to this research shows that participating in surveys is more likely to draw educated people than other groups (Curtin et al. 2000; Singer et al. 2000; Goyder et al. 2002), and also online surveys particularly about genetics have a tendency to draw an educated crowd (Reaves and Bianchi 2013).
Whilst it is not possible to provide robust calculations as to whether the convenience sample gathered via social media is in any way representative of generic users of social media, it does appear that the sample is typical of users of this medium.
Demographics of people who use traditional media
Despite an extensive literature and online search it was not possible to unpick a typical demograph of a person who is likely to respond to a British television news article, let alone one specifically on genetics. However, research from the US shows that viewers of evening news programmes have consistently been on the decline, and this is particularly true of younger age groups (Guskin et al. 2011). The average evening news consumer in the US is over 50, female, with a higher than average level of education and a household income of greater than $75 k and education (Pew Research Center 2012). The sample ascertained via the Traditional Media recruitment method was more likely to be over the age of 41, female and highly educated; this does broadly fit with the profile identified from the American research (which is subtly different from the social media group).
Demographics of people accessed via direct invitation
There is no published publically available data on the demographics of staff approached directly via email listserves to participate in our survey, i.e. from the AGNC, NIHR, Nuffield Council on Bioethics, Wellcome Trust Sanger Institute, Wellcome Trust and Association of Medical Research Charities. However, as a member of the AGNC the first author is aware anecdotally that the majority of genetic counsellors in the UK are female, white, highly educated and aged 31–50. It was not possible to document the demographics of patients who picked up a flyer as part of their attendance at a Science Festival or NHS appointment. What is known, however, is that the demographic data provided in Table 3 largely fits the same demographic data in Tables 2 and 4. It is therefore distinctly possible that the typical demograph of people we have recruited more broadly fits with the type of person who is just generically interested in participating in research about genetics. This leads us to an exploration of the literature already published on attitudes towards various issues surrounding genetics and whether there is a typical profile of participants who engage with this research.
Demographics of people who take part in research about genetics
Research gathering attitudes towards the use of genetic technology have been conducted for over 20 years. Numerous types of participant groups have been sampled and studied; it is difficult to know whether there is a particular type of person who is more likely to be drawn to participate in research on genetics, but it is possible to explore the research that has been done and the socio-demographic data attached to the participants involved. The following studies are very typical examples from an enormous body of literature.
Kerath et al (2013) explored the beliefs and attitudes of members of the public towards participating in genetic research. The survey was distributed to a convenience sample of people attending a network of 15 different hospitals around New York. The sample supposedly represented the ‘diverse, geographic, socioeconomic and ethic catchment areas of the Health System’ (Kerath et al. 2013). Within their final sample (n = 1,041), the majority who chose to complete surveys were over 40, female, white, had a degree or graduate degree, were married and had children.
Cherkas et al. (2010) gathered British attitudes towards personal genome testing from 4,050 members of the public. Their survey was distributed to a convenience sample of twins participating in the TwinsUK Adult Twin Registry, who had been ascertained from the general population. The mean age of participants in the study about genetics was 56, 89 % were female, 79 % had children and the majority were of higher socio-economic status (Cherkas et al. 2010).
Morren et al. (2007) explored attitudes towards genetic testing amongst patients with chronic disease in The Netherlands. The survey was mailed to a nationwide representative sample of patients with chronic disease and returned by 1,496 participants. Within the final sample, the majority of participants were over age 45, 58 % of them were female, 75 % married/cohabiting and 54 % had an ‘intermediate’ or ‘high’ level of education (Wilde et al. 2010).
Whilst there are clearly numerous research projects on attitudes towards various issues in genetics that have been particularly focussed on gathering the views of men (Quinn et al. 2010), certain ethnic groups (Murphy and Thompson 2009, Ahmed, Ahmed et al. 2012) and specific ages of people (Donnelly et al. 2013) these are by far in the minority of the whole body of published work available.
When exploring the literature on the profile of nonresponders to surveys, an interesting Faculty paper was uncovered from William G Smith (2008) at the San Jose State University. Smith summarises the literature on the typical profiles of people who take part in survey research (Smith 2008). He showed that generally people who are educated and affluent are more likely to take part than less educated and less affluent people (Curtin et al. 2000; Singer et al. 2000; Goyder et al. 2002); women are more likely to participate than men (Curtin et al. 2000; Singer et al. 2000; Moore and Tarnai 2002) and white people are more likely to participate than other ethnic or racial groups (Curtin et al. 2000; Groves et al. 2000).
Therefore, the convenience and snowball sample that we have obtained via the three recruitment strategies broadly fit the samples that have been recruited for other research on genetics. The sample also fits with the profile of respondents who generically respond to recruitment invitations to participate in social sciences research. Separate publications will follow that will explore how socio-demographic data are linked to attitudes towards sharing incidental findings from genomics.
Future social science research on genomics could very usefully employ selective sampling frames that specifically target non-white audiences, men, as well as people who have lower educational achievements and affluence. It is only with the contribution of these other groups that useful conclusions can be more broadly drawn on attitudes towards the use of genomic technologies.