Reliability of Measures of Stress (Part B)

June 2011 - (b) Assess the reliability of methods of measuring stress (15)

Firstly, define the key terms in your introduction and outline the methods. You might also want to list some of the issues surrounding reliability.

"Stress is the body's response to when physical or perceived demands on an individual do not match their perceived or actual abilities to deal with a situation. It can manifest itself both physiologically and psychologically, and thus the most common approaches to measuring stress are physiological measures such as GSR, and self-report measures such as the Holmes and Rahe SRRS. However, the reliability of these measures is questionable, because of the changeable nature of stress, issues surrounding bias and standardisation, and confounding variables which mean that measures of stress may not always produce consistent results when the measure is replicated."

There's a lot of ways you could tackle the main body, such as doing one paragraph on internal reliability and one on external reliability, or going through different issues, but I prefer to go through measure by measure, and assessing each one's reliability. For example:

"One prominent measure of stress is the self-report method, which involves asking individuals to record either their stressors, such as their hassles and uplifts, or stress responses, as in the Bradburn Morale Scale. Whilst self-report appears to be an appropriate measures of psychological stress responses as it asks them about how they feel and what has bothered them, it is affected by many confounding variables that mean it's often not reliable. For example, two people may actually have the same number of hassles and uplifts, but individual differences such as the extent to which dishonesty, their mood, their memory, social desirability bias or interpretation affects their responses may result in them self-reporting different numbers of hassles. Thus, the measure may lack both internal and external reliability as well as internal validity for this reason. However, not all self-report methods lack reliability, and those with checklists or which ask closed questions may give more consistent results. For example, Holmes and Rahe's Social Readjustment Rating Scale (SRRS) involves a checklist, which is standardised and consists of the same life events. There is very little interpretation involved in answering the checklists which removes this confounding variable, and as stressors are listed in front of the individuals, poor memory is also less likely to affect the results. Thus, the measure is likely consistent over time, but may still have poor internal reliability due to the impact of mood and bias. Self-report measures are therefore variable in their reliability."

This paragraph includes examples, and a number of factors affecting reliability including standardisation, mood and subjectivitiy. It gives both strengths and weaknesses of the measure, thus fulfilling the "assess" part of the question, too. The result of my answer would consist of:

>> an assessment of physiological measures. I'd include objectivity as a strength, and that it's less affected by the user, but then say it's reliability is damaged by the fact that everybody experiences stress differently - some get headaches, some get muscle tension, others have increased heart rates, some do not experience it very much in terms of biology but their concentration lapses, etc. Also, different factors affect the physiological measures, such as in Geer and Maisel's study, GSR may not have merely measured stress response but other emotional responses, such as fear arousal or shock, and that heart rates as a measure can be affected by general health condition and exercise as well as weight, fitness, etc. 

>>briefly mention combined approach as seen in Johansson. Likely to share strengths and weaknesses of both, but may cancel out the influence of individual differences in terms of stress responses. 

>>Conclusion: both are replicable and thus should have external reliability, though stress is variable and thus the extent to which any measure can be considered to be consistent is questionable as it relies on a large number of variables to be controlled.