Several studies have attempted to determine the role of stress on our health.  Selye’s General Adaptation System argued that although the stress response triggers our immune system, long-term stress leads to an exhausted immune system that leads to health problems.  There are several limitations to research on stress including difficulties in isolating and controlling variables, reliance on self-reported data and limitations in generalization and replication.

One limitation of stress research is how to operationalize and isolate variables. A natural experiment carried out by Kiecolt-Glaser looked at the effect of both stress and loneliness on one’s physical health.  The sample was made up of medical students who were about to take exams.  They were asked to self-report their level of stress and loneliness through questionnaires.  Their level of NK cells was measured before and then after the exam period. The results showed that students with higher levels of stress and loneliness had lower levels of NK cells, or, as predicted by Selye, a weakened immune system.

This study is typical in its limitations. Since it was done under naturalistic conditions, the researchers had no control over the IV and thus a true cause and effect relationship cannot be determined.  In addition, there may have been extraneous variables that compromised their immune system besides the two variables measured; for example, exposure to a cold virus or flu at the university.  Finally, there is the question of operationalization.  Students were asked to assess their own level of stress. This is not a precise measure of one’s level of stress.  Although it is true that they measured NK cell levels before and after exams, the self-reported measure of stress brings any correlation into question as reporting stress maybe socially desirable as a medical student. In addition, measuring one’s own sense of loneliness on a cross-sectional survey is problematic. One’s mood on the day of the survey could have affected their report.

When studies are done under natural conditions, it is not only difficult to isolate and control variables, but it may also be difficult to replicate the study to establish reliability.  The Whitehall study by Marmot was a longitudinal study of the British civil service – a hierarchy. Marmot found that those that were at the bottom of the hierarchy were more likely to suffer from health concerns.  This study used a very large sample size and the participants were matched for many different variables – including smoking and family history.  They found that smokers at the top of the hierarchy had fewer health problems than those at the bottom.  This study reflects Sapolsky’s findings in animal research.  However, it is difficult to replicate a large, longitudinal study of this nature. The British civil service is rather unique in its structure.  Also, the UK has a public health care system.  This makes the study difficult to generalize to systems like in the US.  The study was also reliant on self-reported data, although medical records were used to confirm much of the reporting. When data about health is reported there is always the danger of optimism bias, where people tend to believe that their health was better than it actually was.

Finally, there are ethical considerations in health research. In Cohen’s study he first rated people on their level of stress and then exposed them to a cold virus.  He found those that had higher levels of stress were more likely to develop a cold. Although this could be seen as more direct evidence of the effect on the immune system, it also brings up ethical concerns. Even though informed consent was given, there could be complications from a cold.  Manipulating a study to induce a health problem is ethically questionable.  However, in this case, it could be argued that it was not undue stress or harm because we are exposed to such viruses on a daily basis. However, needing to meet ethical standards limits what a researcher can do with regard to inducing stress or exposing participants to potential health problems.

Strong health studies are prospective, naturalistic and longitudinal.  Not only are these studies not easy to control and difficult to replicate and generalize, they often leads to participant attrition.  Participants drop out of the study for a variety of reasons over time, which may end up skewing the data by only leaving a certain type of participant.

Stress and its effect on our health is complex.  It involves physiological, cognitive and sociocultural factors that interact to lead to health issues.  The inability to directly measure stress and having to rely on self-reports makes the research imprecise and potentially biased. However, the body of research done on this topic has shown us trends which have led to improved approaches to health and wellness.