With her wealth, status and access to hot IPOs of the late 1990s, style diva Martha Stewart was no average investor.

But a professor of behavioral finance contends in a recent study that Stewart succumbed to an investing weakness that is far from extraordinary: a reluctance to sell either winning or losing stocks.

Meir Statman, professor at Santa Clara University in Santa Clara, Calif., reviewed brokerage account records entered as evidence in court as well as testimony in the trial.

Stewart, who founded

Martha Stewart Living

(MSO)

and worked as a stockbroker early in her career, was sentenced to five months in prison Friday, having been convicted March 5 of lying to the government about her well-timed sale of about 4,000 shares of

ImClone Systems

(IMCL)

in December 2001.

Here's a look at the stock trading the study analyzed.

On Dec. 20, 2001, her personal account at

Merrill Lynch

(MER)

consisted of 25 stocks valued at $2,385,289. Of the 25 stocks, 22 had unrealized losses, including

Amazon

(AMZN) - Get Report

,

Apple

(AAPL) - Get Report

,

Nokia

(NOK) - Get Report

,

DoubleClick

(DCLK)

,

Lucent Technologies

(LU)

and

Kmart Holdings

(KMRT)

. For tax reasons, she sold these by year's end for a combined loss of $704,525.

In an email to a friend on Dec. 22, she wrote: "Just took lots of huge losses to offset some gains, made my stomach turn."

Stewart, according to Statman, was exhibiting that common investor weakness or "cognitive bias" known as framing, "where investors open mental accounts when they buy stocks and close them only when they sell the stocks.

"Investors do not have to acknowledge paper losses, because open accounts keep alive the hope that stock prices would rise and losses would turn into gains," Statman writes. "But hope dies when losses are realized. Realization of losses brings the pain of regret investors feel when they find, too late, that they would have had happier outcomes if only they avoided the losing stocks."

Stewart, like many ordinary investors, was also reluctant to sell her winning stocks.

"While realizing gains brings the joy of pride rather than the pain of regret," Statman says, "realizing gains involves a choice, and choice sets the investor up for regret. How would you feel if you chose to sell a stock only to watch it soar on the following day?"

Stewart would certainly have been better off quickly flipping her shares in the IPO offerings she was allowed to join, such as

Agilent Technologies

(A) - Get Report

; Digex, which was acquired last year by

MCI

(MCIP)

; Palm, now known as

palmOne

(PLMO)

; and others.

In particular, she purchased 350 shares of Palm, for example, at the $38 offering price, which soared to $95.06 the first trading day. It never again reached that peak, but in the months following she could have sold her shares for more than the $3.47 each she accepted on Dec. 21, 2001.

Her broker and co-defendant, Peter Bacanovic, formerly of Merrill Lynch, told

Security and Exchange Commission

investigators that he had tried unsuccessfully to persuade her to take some gains off the table.

"I have often recommended that she sell stocks at high prices to lock in gains," Bacanovic said. "And she's very loath to sell stock -- ever -- and has often watched good gains evaporate.

"The perfect example would be Kmart.

I told her, 'Please, let's sell the stock at $8. You have so many shares, your basis is $7.40. You're getting really close here.' We wound up selling the stock at $5. But I think, even that, in hindsight, looks lovely."

(Ironically, since then the company, now called Kmart Holdings, entered into and emerged from Chapter 11 bankruptcy protection, wiping out the holdings of most shareholders. Its stock, however, has tripled in value in the past 12 months to about $79.)

Maybe that's why the jury found it hard to believe Stewart's rationale for selling all her shares of ImClone Systems, the most profitable stock in her portfolio, on Dec. 26, 2001: that she had reached an agreement in advance with Bacanovic to sell ImClone if its price fell to $60 a share. That would have required a major change of investment philosophy, from that of a normal investor of the greedy '90s to a rational one with a sense of self-preservation.