I’ve been working for an urbanism agency for over a year now. For the most part, my job consists of conducting surveys. I am currently working on a new contract that we signed a few weeks ago. The city of Montreal asked us to evaluate their clients’ level of satisfaction concerning the service that they received during their latest demand for a construction/renovation license.

As I was making my way down the call list last week, I realized a pattern was taking place. Depending on the time at which I reached the clients, their answers to my questions greatly varied.

I called people at three specific moments in the day when the chances of people answering their phones are greater: right after lunch time (about 1pm), right before supper time (5:30pm), and right after supper time (7pm).

Did you notice a theme?

Yeah, food.

Here is a brief overview of the type of questions found in the survey.

I noticed that clients who had just eaten consistently rated their level of satisfaction as higher than those who hadn’t eaten yet. Does food deprivation really affect responses on a survey? It could have been a coincidence so I decided to look for theories related to my hypothesis.

I found that serotonin is decreased in food deprived individuals. Low levels of serotonin are also associated with aggression and impulsivity. Furthermore, a study found that “when serotonin levels are low, it may be more difficult for the prefrontal cortex to control emotional responses to anger that are generated within the amygdala.”

Perhaps, this could explain why the clients who hadn’t eaten yet reported being less satisfied. Because their serotonin levels were low, if they had encountered any difficulty during the process of their license demand, they were less likely to control their emotional responses to anger, and more likely to let that anger influence their other responses in the survey. On the other hand, the clients who were not food deprived may have been able to comment on the problem they were faced with while still being able to separate it from their overall level of satisfaction.

It seems, then, that food deprivation does affect responses on a survey. Should I be concerned about the reliability and validity of my results?