Imagine you had two applicants for a graduate vacancy. Both are well-qualified, with strong academic credentials, plenty of work experience and a passion for advertising. How would you choose between them?
Perhaps you would turn to less consequential factors such as their hobbies or their dissertation topic. But what if these too were identical?
Imagine, in fact, that the only difference was their name. Who would you rather work with, Daniel or Danesh?
Everyone I have spoken to in advertising claims that a name wouldn't influence their choice, whatever ethnicity it signified. But claims are an inaccurate guide to behaviour - few people would publicly admit to being unethical, let alone discriminatory.
Why, then, is there a perceived lack of diversity in the industry?
In order to understand how ethnicity affects a candidate's chance of getting an advertising job, we devised a simple experiment. Along with my colleague Claire Linford and one of our work-experience students Hannah Jones, we created a series of fake graduate CVs and covering letters, with strong academic results from prestigious universities and relevant work experience.
We then made two versions of each CV: one with a white-sounding name (such as Daniel Hopkins) and the other with an Asian-sounding name (Danesh Bhat). There was absolutely no difference between the CVs bar the name. We had eight templates in total, featuring male and female candidates.
We then applied for graduate jobs at about 40 top UK media and creative agencies. Half received the CVs with a white-sounding name, half the ones with an Asian name. All in all, we sent out 294 CVs.
Encouraging results
Here are the headline results: both sets of applications had a 23 per cent response rate. They received a similar number of interview offers: 7.4 per cent for the "white" candidate and 6.9 per cent for the "Asian" one.
This is encouraging as many industries are blighted by discrimination.
The Department for Work & Pensions conducted a similar experiment in 2009 across a variety of industries. It sent out 2,971 CVs and found that the white candidate was offered an interview 10.7 per cent of the time, compared with just 6.3 per cent for the ethnic-minority candidate. White candidates were therefore 74 per cent more likely to be offered an interview than ethnic minorities.
Unfortunately, high levels of discrimination seem to be the norm in many industries. The original experiment carried out in 2004 by Marianne Bertrand and Sendhil Mullainathan, which was published in The American Economic Review, investigated discrimination in Chicago and Boston. It discovered that white-sounding names were 50 per cent more likely to get a callback after a job application than those with an African-American name.
Education is key
So if agencies are not discriminatory in their hiring practices, what explains advertising's diversity issue?
According to the IPA, 89 per cent of agency employees are white. However, this sounds less shocking when you realise that 87 per cent of the population are white, according to the Office for National Statistics.
Advertisers recruit from around the country, so you would expect these figures to be close. But why the two-percentage-point lag?
Our experiment only investigated discrimination at the interview stage of the job-seeking process. It may be that discrimination occurs later, perhaps when agencies offer jobs. Or it may be that fewer people from ethnic minorities apply for jobs in advertising, possibly discouraged by its image. If this is the case, then we should promote our finding that advertising is a more welcoming sector than others.
While our experiment shows that our industry is far less discriminatory than others, it highlighted other areas of concern. The most striking point is how hard it is to get a job in advertising. Only 7 per cent of applicants were offered an interview and, less forgivably, only a quarter of candidates even received a reply to their application. Considering that an agency's success relies on having the best talent, this is a short-sighted approach.