Tootsie Roll Industries Inc. posted $496 million in sales last year. At Kraft Foods Inc., sales totaled $42.2 billion, about 85 times as much. Yet Tootsie Roll considers Kraft a "peer" when deciding how much to pay its top executives.

That comparison may be extreme, but new academic studies suggest many companies include bigger and more complex rivals in compensation peer groups used as a factor in setting pay, helping to boost executive paychecks.

Critics long have said that directors and compensation consultants skew these comparisons to push pay higher. But the extent of the phenomenon was unknown until the Securities and Exchange Commission in 2006 began requiring companies to disclose the members of such groups.

One study, led by University of Maryland Prof. Michael Faulkender, found that companies tend to select peers with highly paid executives. The study, of 657 companies, suggests "that firms are gaming the selection of peers for the purposes of gaming executive compensation," Mr. Faulkender says.

Companies with what experts consider weak corporate governance -- where the CEO is chairman of the board or where directors serve on multiple boards, for example -- are more likely to choose highly paid peers, the study found.

Compounding the problem, roughly 40% of companies specify that they aim to pay their CEOs more than the median of their peers. As a result, pay "continuously spirals upwards," says Carol Bowie, director of the Center for Corporate Governance at RiskMetrics Group Inc., a proxy adviser.

A second study, led by Boston University Prof. Ana Albuquerque, compared the peers chosen by more than 1,000 companies against "expected" peer groups chosen by a computer, based on a company's industry, size, performance and other factors.

The study found that companies tend to select peers with better-paid CEOs. On average, the authors estimate that the bias boosts pay about 5%, or $340,000 annually.

The authors say the skewed selections may be justified in some cases, by allowing companies to recruit and retain well-regarded CEOs. "I don't honestly believe that everybody is self-serving," says Rodrigo Verdi, a co-author who is a professor at Massachusetts Institute of Technology's Sloan School of Management. "But I also don't believe that everybody is doing this purely naively."

Wayne R. Guay, an associate professor of accounting at the University of Pennsylvania's Wharton School of business, says he finds the papers "comforting" because they suggest that peer-group selection is a small factor in CEO pay. Looking at peers is the only "feasible" way for directors to gauge how much to pay a CEO, he says.

Charles Elson, head of the Weinberg Center for Corporate Governance at the University of Delaware, says the studies could add fuel to debates among lawmakers and regulators over two executive-pay issues: the independence of board compensation committees and the role of pay consultants, who often choose peer groups. Peer groups are "easily manipulated," he says.

Consider International Game Technology, a Reno, Nev., maker of slot machines and related software. IGT posted revenue of $2.6 billion in the fiscal year ended Sept. 30, 2007. Most of its 20 peers that year were software and casino-equipment companies with annual revenue of $1 billion to $4 billion.

But IGT also included casino giants Harrah's Entertainment Inc. and MGM Mirage, as well as mailing-equipment maker Pitney Bowes Inc. Those companies posted 2007 revenue of $10.8 billion, $7.7 billion and $6.1 billion, respectively.

Median CEO pay among IGT's peers in 2006 was $10.8 million, according to Ms. Albuquerque's study. But the 20 peers chosen by the computer model -- which included fewer big companies and more specialty-equipment makers -- paid their CEOs a median $2.2 million. IGT paid its chairman and then-CEO, Thomas Matthews, $7.5 million in fiscal 2007.

IGT Chief Financial Officer Pat Cavanaugh says the peers were chosen to reflect the breadth and complexity of IGT's business, which spans equipment, services and software. "If we're going to attract a CEO...it's going to be someone who either has [run] or is capable of running a very large organization," he says.

All 12 of Tootsie Roll's peers in 2007 posted higher revenue than the candy maker, including food giants Kraft, General Mills Inc. and Dean Foods Co. Median CEO pay at those companies was $5 million the prior year, according to the Albuquerque study.

The computer chose 12 "expected" peers, replacing the big companies with smaller specialty-food makers; median CEO compensation at those companies totaled $1.3 million.

Tootsie Roll paid CEO Melvin Gordon $2.5 million in 2007. Tootsie Roll President Ellen Gordon, his wife, says the chosen peers "may be viewed as competitors of Tootsie Roll from a product and executive talent standpoint." She says a compensation consultant helps the company adjust the pay data "to be more relevant for a company with the market capitalization of Tootsie Roll."

Write to Cari Tuna at cari.tuna@wsj.com

Printed in The Wall Street Journal, page A3