A poster warning New York subway riders to avoid consuming too much sugary soda — by picturing a diabetic amputee — gained notoriety last week because, it turned out, the man pictured wasn’t an amputee at all. He’d been photoshopped.

But the ad should have generated controversy for an entirely different reason: It’s the latest inappropriate use of a scary, graphic image to try to produce a government-mandated change in consumer behavior.

Though fear is the weapon of choice for government health police — whether they’re trying to stop people from eating fatty foods, smoking cigarettes or drinking too much — it simply doesn’t work.

The case for alarmist warnings is based on four assumptions. First, people wish to avoid disease and death. Second, consumers suffer from an “information deficit,” that is, they either don’t understand the risks of a given behavior or they underestimate those risks. Third, once they know that a certain behavior or product can lead to disease and death, they will avoid it. And, fourth, warnings give people the information necessary for them to change their behavior.

The reality, however, is that assumptions two, three, and four are, for many people, false.

Many people filter out much of the information available to them because they find it neither relevant nor interesting. Also, warnings aren’t processed because individuals tend to avoid information that has negative self-implications. Through a process known as “cognitive readjustment,” people tend to exempt themselves as individuals who should be concerned with a warning. Even though someone has read and remembered a warning, they also can discount its personal applicability.

Even warnings that are read and processed are often discounted due to what experts call “warning fatigue,” where the overabundance of warnings or the familiarity of a specific warning diminishes its effectiveness. Consequently, the empirical research on scary tobacco warnings, for example, turns conventional wisdom on its head. Such fear-based public health education campaigns simply do not work.

The evidence from empirical studies of their effects in real world settings demonstrates these types of warning failures are extensive. Almost a decade after the federal government mandated warnings on alcohol products, neither the risk perception nor the drinking behavior of those drinkers most likely to be a risk to themselves or others had changed. There is equally compelling evidence about the failure of food labeling. The USDA found that labeling is an ineffective policy tool.

The truth is that more often than not, scary or detailed warnings cause many consumers to disregard the information completely. A consumer’s income is the key factor in determining which foods, for example, are purchased, and that income cancels out the effects of information.

The danger, however, is not simply that labels and warnings will fail; they may also be counterproductive. For example, large numbers of excessive risktakers display what psychologists call “reactance,” in which there is a high level of resistance to the demands of outside authority and control.

For these individuals, a warning label represents an attempt to unreasonably (at least from their perspective) shape their behavior and makes them more likely to ignore rather than heed the warning. Warning labels also highlight risk, and for those attracted to risk-taking, this serves to make the very thing warned about more, rather than less, attractive.

The use of such warnings contradicts two of the central principles of medical ethics and the ethics of health promotion: beneficence — doing good — and nonmaleficence, avoiding harm where compelling evidence demonstrates that warnings will do no good and might cause harm. Curiously, the FDA now advocates scary tobacco warnings even though its own study found them ineffective.

Today’s health warnings improperly utilize the state’s legitimate authority in a manner that converts public health advertising into displays for the government’s opinions.

New Yorkers have the right to shape their own lives, together with the responsibility for the results of doing so, without the intrusion of Mayor Bloomberg’s taxpayer-funded advocacy.