Does being hungry or thirsty alter your morality?

Does being hungry affect morality?


I am a big fan of Dan Ariely — the behavioural economist and author of a number of books on human behaviour such as “Predictably Irrational”.

I was recently listening to one of my favourite interviews with Dan conducted by David Kadavy (check out his podcast here) and it inspired me to cover one of Dan’s studies.

We are all probably familiar with the ethical dilemma of the starving person stealing food to eat.

It seems to be a staple in fiction and philosophy (e.g. Les Miserables).

Many people would not think it immoral to steal food if one is starving. However stealing food in other circumstances would be considered immoral.

It would seem that a visceral state (in this case hunger/starvation) makes a normally immoral action acceptable (at least objectively to others).

What about in a subjective situation? Do such states actually alter people’s morality and actions?

This paper looks at how visceral states (like hunger, thirst etc.) affect morality, specifically by seeing how likely people are to cheat to win a hunger or thirst quenching prize.

I will present my TLDR style summary first with the more detailed discussion following afterwards:

The Summary

In an elegant set of 3 experiments the authors showed that people are more willing to cheat in a task to win a food or water prize when they are in a hungry or thirsty (visceral) state.

In each case the instances of winning were substantially beyond the 50% wins one would expect from chance, ranging from 68–80%.

Further the cheating did not seem to correlate with the perceived economic value.

This suggests that the cheating and reduction in moral behaviour applies solely to items that will immediately fulfill the visceral need.

To apply this to the often quoted moral example, someone who is hungry is more likely to steal bread, but no more likely to steal something else of equivalent value.

Visceral drives do seem to affect morality but only in ways which serve to directly fulfil them.

The Study Comprised Three Experiments

The study was split up into three separate experiments. Each one was slightly different so I will cover them separately because of this:

Experiment 1


This involved 144 students, staff and visitors at a University of Florida campus food court at lunch time.

The aim was to use a time and location where there were likely to be a substantial number of hungry (as well as non hungry) participants.

They were told they were taking part in a marketing survey about a snack pack (containing potato chips, a granola bar, Snickers bar and a mint).

Participating in the survey would enable them to win one of these packs but there were not enough for everyone to get one.

They were asked a variety of questions about mood and physical state (hunger/thirst etc) as well as how much they would be willing to pay for the pack.

In order to win the prize they had to roll a die in a cup and self report if they won or lost without the ability of the tester to see if they lied.

Rolling an even number resulted in a win, an odd number was a loss.


If the results were purely dictated by chance and people were reporting honestly we would expect 50% of participants to roll a winning number.

In this case 76.5% reported winning and the greater the hunger the greater the probability was that they would “win”. That is to say that being hungry increased the likelihood of reporting a winning number.

This suggests that being hungry makes people more likely to cheat than they would if they are not hungry.

Did Economic Valuation Change?

A smaller association was found between the price people were willing to pay for the pack and cheating.

This suggests that the overall monetary value that they applied to the pack was less of a driver to cheat than the hunger state itself.

The implication is that hunger increases cheating in order to satiate the hunger itself rather than for a financial/monetary gain.

Experiment 2


This study looked at thirst.

Exercise is a good way to induce thirst and so the experimenters approached people entering and exiting a gym.

Those entering the gym would likely be less thirsty than those leaving and hence would act as a control population for the task.

62 participants were used (Cornell University Students).

They were asked to fill in a questionnaire on mood and visceral states, assign a price to a bottle of water and then think of a number between 1 and 10 to win one of the bottles.

They were told that there were not enough bottles to go round and were informed after thinking of a number (but before saying what it was) that they would win if they guessed an even number.

This gave them the opportunity to cheat by lying about the number they thought of.


As expected participants were thirstier after working out, with a mean score of 2.45 (out of 4, where 4 = extreme thirst) vs a mean of 1.63 before working out.

Thirsty participants “guessed” a winning number 80% of the time as opposed to the 50% one would expect due to chance.

So once again people in a visceral state, in this case being thirsty, were more likely to cheat to win a prize that could quench that thirst.

Interestingly those who took part before working out had a rate of 40% which is less than what one would expect from chance.

Even vs Odd Bias?

The authors suggest this may be due to people having an inherent bias for selecting odd numbers (suggested by past research).

To investigate this further they re-ran the experiment (with 72 subjects) swapping the winning numbers from being even to odd.

In this case the winners were still very similar in the thirsty group at 78.9% but increased to 56.9% in the pre-workout group.

This would suggest a bias for selecting odd numbers, however, it does not affect the outcome.

Financial Value vs Thirst Quenching Value

Again the authors wanted to know if this had any relationship to the price they assigned i.e. was this purely to quench thirst or was it a general decline in morality?

In this case thirsty participants did not assign a higher value to the bottle of water.

The authors describe this as a null result because they expected them to assign a higher monetary value.

In the first experiment there was a small but significant difference — not enough to account for the level of cheating but it was present.

This suggests that the desire to cheat is motivated purely by the desire to quench thirst and not by any perceived increase in financial benefit.

Perhaps the viscerally derived value of the prize is considered or conceptualised separately from the cognitive, economic value.

Of course it is possible that one of these results (first or second experiment) represent an anomaly as it relates to the economic valuation.

It is also possible that there is a difference in how the two different visceral states alter judgement. Neither of these possibilities are sufficient to negate the rather large correlation between cheating and visceral states.

Experiment 3


The third experiment sought to examine and clarify the issue of economic vs visceral reward value.

The same basic method of manipulating thirst and using gym attendees from experiment 2 was used.

However in this case 91 people were split between being offered a bottle of water or a souvenir pen (of the same value).


Thirst was again found to be higher in post workout subjects with a mean of 2.48 vs 1.77 pre workout.

Before working out 45% of people won the water (slightly below chance), whereas 60.9% won the pen.

This suggests people were cheating to win the pen when they were not thirsty.

Cheating to Lose?

People who were thirsty (after working out) won the water 68% of the time and interestingly only won the pen 30.4% of the time (half the value for the non-thirsty group).

This finding suggests that thirsty people actively “lose” when the prize is the pen despite the financial value being the same.

It also confirms that cheating to win in the thirsty state is only related to winning items which will directly affect that state.

They actively “cheated” to lose the pen even though the financial value was the same.

This suggests a lower overall value being attributed to the pen in the more thirsty state in spite of this.

Moral Virtue?

Although the authors talk about this as being cheating I think this needs needs to be distinguished from a moral perspective.

Whilst still a form of dishonesty, one could say that cheating to lose could be considered a more morally “virtuous” form of cheating.

To me this result suggests that moral standards only decline in relation to the prize that will satiate the visceral drive (i.e. thirst) which is consistent with the previous experiment.

Further Discussion (all 3 parts)

The results of the study overall are interesting because they suggest that morality can be affected by visceral states such as hunger and thirst.

This makes some implicit sense from an evolutionary perspective.

Just as interesting is the point that the decline in morality only seems to pertain to items that might immediately quench the visceral state.

Inconsistent With Drug Addiction?

This makes sense but may be inconsistent with certain other types of criminal behaviour:

It is known for example that drug addicts will often steal items to sell and buy drugs.

This would seem to contradict the findings here in that in that particular case the immoral behaviour does not result in a direct fulfilment of the visceral drives.

The authors suggest that this is because other cognitive mechanisms may be involved in that process resulting in greater premeditation.

I think there may be a simpler answer though.

In the case of drugs, whether they are legal or illegal, they are much more difficult to get hold of and hence steal than a loaf of bread or bottle of water.

Prescription drugs of abuse like opioids are kept under tight security. Illegal drugs are not as widely available and are likely to be kept just as secure by those who are selling them.

It is likely that drug addicts are forced to take the extra steps of stealing and selling other items because the option of stealing drugs directly is not available to them.

Multipart Design

Overall I think this study is well done and presented.

The way the experiments were split up into three parts is quite elegant and suggests a good way of keeping things simple and breaking down the effects being studied.

It also helps to reduce confounding variables which I will discuss later.

Sample and Resource Limitations

As most people will be aware psychology gets criticised a lot for being a soft science and unfortunately there are limits to how much can be done to tackle this.

Dealing with human behaviour and motivations is by definition more complex than simply measuring a variable like the voltage in a wire.

It is also constrained by the generally high resource and time cost that comes from the involvement of people.

Ideally larger sample sizes would help to increase the validity but that would also hugely increase the cost.

It would likely be more viable to collate the results of multiple studies similar to a meta-analysis in medical research.

As I have discussed before this is not as good as a properly designed single study of equivalent size but it does give you evidential validity which is somewhere between the two.


I would be remiss if I did not mention the recruitment which involved a specific population i.e. people on university campuses.

Is it possible that these people are in some way different and have different responses to other parts of the population?

I think it unlikely but the only way to know for sure would be to compare with other groups by repeating in different locations.

Confounding Variables

Another major area of concern with psychology (and medical research) is that of confounding variables.

I think here the main potential confounding factor of the economic value of items is well addressed both by splitting the experiments up and using extra measures.

Where more than one choice was available the financial value was equal. In other cases the economic valuations between the different groups were compared.

This helps to illustrate that the change in morality was specific to items that would reduce hunger/thirst rather than other items.

This reduces the likelihood that other motivations were involved in the change and also suggests that overall morality is not impaired.

Very Well Written and Presented

Unlike a lot of studies that I read this one is really well presented in terms of how easy it is to read and follow.

I don’t think it is a coincidence that Ariely is a best selling and thoroughly enjoyable writer.

Obviously papers are rarely written by one person and in this case I think all the authors should be commended on this.

Good communication is sadly one of the most neglected aspects when it comes to presenting research.


This is a fascinating and well presented study which suggests that morality is indeed altered by visceral states.

Further the alteration is not necessarily a global change in morality but only favours outcomes which directly satisfy the visceral state such as hunger.


  1. Williams, Elanor F., David Pizarro, Dan Ariely, and James D. Weinberg. 2016. “The Valjean Effect: Visceral States and Cheating.” Emotion 16 (6): 897–902.

Thank you for reading

Want more like this? — Follow me on Twitter or connect on LinkedIn to keep up with my latest posts. The original version of this post was made on my Steemit blog here (Steemit is a new way to monetise your content find out more here).

All uncredited images are taken from my personal Thinkstock Photography account. More information can be provided on request.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.