ANALYSIS: Salaries - Sex, lies and multiple regression analysis

There's been an argument raging about sexual discrimination in PR

for many years. Is there a pink ghetto? What happened to the glass


Adam Leyland reports on some new findings

Not so long ago, the upper echelons of Burson-Marsteller were a bastion

of clubbish, male old-school values. Yet as a USA Today picture

demonstrated last month, nowhere will you find a better example of how

the feel of PR has changed. In the picture, the 80-year-old chairman,

Harold Burson, was surrounded by a posse of high-powered Burson


'The biggest change in the four years I've been here is the number of

senior-level appointments just below the executive board level,'

confirms Celia Berke, MD of human resources at Burson. 'They're being

groomed, but not because they're women; it's because they're the best at

what they do.'

Women moving up, or are they?

Indeed, women run almost all of the Burson practices - each responsible

for P&L - including Linda Recupero (marketing); Amy Wadler (healthcare);

Heidi Sinclair (technology); and Judy Mackey (corporate/financial). So

if Burson, along with the entire PR landscape, has changed so radically,

why is it that a recently published book should continue to suggest that

women lag behind men in terms of both pay and opportunity?

A just-published PRSA 'gender study' appears to confirm the problem.

In its sample of 894 members, the report's author, Elizabeth Toth, says,

'We continued to find significant differences in salary which we

attribute to discrimination,' noting a dollars 17,000 discrepancy

between pay for men and women. 'Same as ever,' sums up Kathy Lewton, CEO

of the PRSA, and a 20-year-long supporter of a battle within the PRSA to

eliminate sex discrimination from the industry.

A first glance at the results of the PRWeek/Text 100 Salary Survey 2001

(PRWeek, March 26) suggests the same statistical pay disparity


But as PRWeek found, gender-based salary discrimination requires a

deeper analysis. We asked Jim Hutton, associate professor of marketing

and communications at Fairleigh Dickinson University, and an expert in

statistical interpretation, to analyze the salary data from PRWeek's

sample of over 5,000 PR execs.

Hutton used a technique called 'multiple regression analysis' (accepted

as the legal standard in court cases of gender discrimination) to

examine the factors that drive salaries. As Hutton explains, while there

is no doubt that men in PR earn more money than women, 'to assume that

the salary differential is the result of discrimination is completely


The question is not whether men make more money than women, but whether

the reasons for the difference are legitimate.'

The numbers behind the numbers

In analyzing the numbers, multiple regression analysis calls for an

examination of several other 'reasons' or 'variables' (as many as 10 or

11) that might explain or influence the statistics: experience, age, the

length of continuous service, the type of organization you work in

(agency vs. corporate), the type of PR you practice, the industry you

work in, education, color, etc. The regression software program then

picks out the 'reasons' in order of importance.

The most significant 'variable' that emerges is 'years of experience,'

accounting for 26% of the discrepancy in salaries. Table 1 shows that

while women now outnumber men by as much as 3:1, men have substantially

more experience than women. Table 2 shows that men, on average, work

longer hours than women. This accounted for an additional 4% of the


Of equal importance (4%) was the 'type of organization.' As Table 3

shows, men tend to work in corporate PR, where the pay is somewhat

better. And contrary to the suggestion that men dominate in the flashy

practice areas like crisis management, reputation management, or IR, as

Table 4 shows, this accounts for just 2% of the variation. Age (perhaps

better described as maturity since the effect of experience has already

been examined), also played a relatively minor role, accounting for 2%

of the variation.

And lumping together other less-important variables explains a further


After all these factors are considered, Hutton's analysis shows that

gender accounts for just 1% of the variation in salaries. And as he

explains, even this 1% may not necessarily represent discrimination.

'There are at least three non-discriminatory factors that were not

included (in the PRWeek/Text100 Salary Survey 2001 questionnaire) which

might explain all or part of the remaining gender gap. For example, 'job

continuity,' or 'breaks in service,' was a major factor in explaining

salary differentials in a salary survey of the marketing/advertising


What the statistics really mean

How does this jive with the PRSA's findings? Even Lewton admits that

'agencies have been so competitive, and have had such a need for talent,

that gender literally doesn't matter.' Toth insists that the gender

study results were subjected to regressional analysis of their own. The

report notes that 'statistical tests showed that even when years of

experience, job interruptions, age and education level were accounted

for, gender still made a significant impact on salary.' However, when

asked what the percentage difference was after stripping out these

factors, neither the report, nor Toth herself, were able to specify the

extent of the gap, either in dollar or percentage terms.

And what do other salary experts say? Jean Allen, senior partner in

Heidrick & Struggles' communications group, notes a definite change. 'I

think the pay structure is far more equitable than ever,' says


'Women are reaching the senior levels where they can demand higher

levels of pay.' James Wyckoff, SVP at Marshall Consultants in New York,


'The biggest factor in recent years has been determined by the industry

in which you work, with hi-tech leading the charge.'

However, Ben Long, at Travaille, a Washington, DC-based recruitment

agency, says he still sees sex discrimination in 'subtle ways.' 'It's

done by title manipulation, or bonus structure,' claims Long. 'If a man

replaces a woman, they are more likely to pay what the outgoing person

was paid; if the incoming person is a woman, I'll often see them coming

in at a lower rate.' Still, Long, like others, acknowledges that it's

'slowly getting better. In some sectors, like fashion and cosmetics,

it's not prevalent at all. In aerospace, it's more ingrained. But I'm


Where women may lose out, says Allen, is through their lack of

negotiating skills. 'Men are much more likely to name their price,

whereas women will couch the issue in terms of a request for that

salary. This effect does wash out the further you go up the pay scale,

but it still happens, even at the most senior levels.' Long, however,

sees this changing. 'The women I've been dealing with lately are damned


Of course, it could be argued that the greater proportion of men in

corporations and in practice areas like IR is a form of de facto salary


However, as we have seen, these factors have little impact on pay (some

6%). Besides, there is no evidence that women are being disbarred from

these areas. Indeed, with 37 female CEOs among the top 100 PR agencies;

and women holding top posts at the likes of GE, Procter & Gamble,

Microsoft, American Home Products, Boeing, Home Depot, National

Geographic, Dupont, Lucent, Glaxo Wellcome, Dow Corning, Paine Webber,

Barnes & Noble and Starbucks, should we really be worried?

Have you registered with us yet?

Register now to enjoy more articles and free email bulletins

Already registered?
Sign in