Thursday August 13 2009
The evidence on which these claims are based is weak
“In times of crisis, there is nothing like a nice cup of tea,” according to the Daily Mail, which says the drink is helpful when trouble is brewing. The newspaper quotes scientists who believe it is not just the drink, but even the simple act of putting the kettle on that can stop stressful situations from boiling over.
The news is based on a scientific study commissioned by the Direct Line insurance company. The researcher took 42 men and women and gave them a mental task to complete to a tight deadline. Half were then given tea, while half were given water. Those who drank water were 25% more stressed after the task than before. However, the tea drinkers were actually calmer than when they began the test. Focus group research also found that “a stressful day at work” and “office politics” were the most common types of daily dilemma solved by tea.
While news on this study has created a stir, the evidence on which any claims are made is weak. This was a small study with several limitations that has not been peer-reviewed. This means that there is low confidence in the accuracy of the findings. In particular, it appears possible that all participants and researchers not only knew who was conducting the research and why, but also which group the participants had been allocated to. Also, it is not possible to say how precise the reported figure of 25% is. Although some people may find that tea helps them to relieve stress, this study should not influence tea drinking or kettle-use habits.
Where did the story come from?
Dr Malcolm Cross and Rita Michaels, psychologists from the City University London, conducted this research. The study was directly commissioned by Direct Line Insurance plc and published in summary form on its ‘tea-mergency’ website, a website that claims to have been “solving insurance related tea-mergencies great and small with a dedicated team of customer service and claims management experts”. The website does not say whether this research has been reviewed by other experts in the field to ensure its reliability.
What kind of scientific study was this?
This was a non-randomised controlled study in which the researchers aimed to measure and understand how well tea calms people during and after an episode of anxiety. They also wanted to see how the social and psychological ritual of tea-making affects stress.
The researchers selected 42 participants (21 men and 21 women) and divided them into two evenly-sized groups: a tea group and a non-tea group. They then administered two widely used, validated psychological tests to measure anxiety, called the Spielberger and the State-Trait Anxiety Inventory for Adults (STAIA). The STAIA uses scales to measure both the state of anxiety of a person at the time of the test and their ‘anxiety trait’, which is their tendency to be anxious.
The participants were given a stressful timed test in the form of a mental attention task, designed to be challenging and to induce anxiety. The participants were asked to cross out or leave the letter ‘d’ in a passage of text depending on whether it had either two marks above it, two marks below it, or marks above and below it. The letter ‘p’ was also interspersed in the text, both with and without marks above and below, making the test quite difficult to perform under the pressure of a time limit.
Members of the tea group were then given tea, while those in the non-tea group were given a glass of water. They were once again given a STAIA test to measure their anxiety levels after completing the task. The time between the first measure of anxiety, the stress-inducing test and the second measure of anxiety were the same for both groups.
In focus groups after the end of the study the researchers asked the participants structured questions about the emotional significance and impact of making tea, what tea meant to them, what it made them feel and why.
What were the results of the study?
Although the groups showed no significant difference in anxiety prior to the testing, the researchers say that after the study there were significant and marked differences.
The researchers say that the post-test level of anxiety in the tea group was lower than that of the non-tea group, with the non-tea group showing a 25% increase in anxiety levels after the stress-inducing task, compared to a 4% decrease in anxiety levels in the tea group.
The researchers say that, “Focus groups conducted with participants after the experiment confirmed that the ritual of making and consuming tea does make an important contribution to the overall effect of mediating stress”. For example, one participant said that tea created a “chill-out moment” that helped to “draw a line under” their stressful experience. Others reported feeling “looked
after” and “cared for” by having tea made for them.
What interpretations did the researchers draw from these results?
The authors conclude that tea not only reduced anxiety caused by the task, but that it also lowered the group’s anxiety below the starting level, effectively making participants more relaxed than they were prior to the stress-inducing activity. Conclusions from the focus groups were that, during periods of stress, “tea’s reputation for inducing calm extends beyond the effects of its physical properties on our bodies and brains”.
What does the NHS Knowledge Service make of this study?
It is hard to disagree with the researchers who say, “most of our mothers would tell us, if you are upset or anxious, put the kettle on”. However, research should not be designed merely to support current beliefs. There are several aspects of this study and its reporting that mean the results do not support its conclusions:
- The researchers do not report sufficient detail on the participants, for example how the sample of 42 were selected, what age they were, where they come from, whether they like tea or not, whether mental illness had been excluded and what time of day the tests were conducted. Uneven distribution of these characteristics between the groups could account for the differences seen.
- The method of allocation to the groups is not described. This is important as, if this was a non-randomised process, it is also likely that the groups were different before the study started. Examination of the graphs provided shows that the groups differed by almost 100 points on the anxiety scale before the experiment began.
- No information about the statistical significance tests is reported. This would have allowed an independent review of how robust the effect seen was. It is only from the graph that the suggestion of a difference can be inferred.
- There was no blinding in this study, meaning that all the participants and testers were aware of the drinks being consumed. It is very likely that the participants knew what the researchers were interested in finding before they completed the anxiety testing, even sub-consciously this could have affected their responses.
Overall, this study seems to be soft evidence supporting a common-sense theory. Better designed trials will be needed if anyone is really interested in how tea helps to calm tea drinkers.