In this essay the research area of interest is the effect of the social environment on anxiety. Two different methods are examined to see how they have contributed to this research in two particular studies. The methods are: meta-analysis arising from self-report data and semi-structured interviews.
The first study to be examined is Twenge (2000). Twenge found that anxiety had increased considerably over the time of the study and the environmental attributes of low social connectedness and high threat explained these high levels of anxiety. The method used in the study was meta-analysis. Meta-analytic techniques were used to gather and analyse data from samples of American college students and children between the years 1952 and 1993. The participants completed self-report measures of anxiety and neuroticism. For the college students the initial research was conducted using the Taylor Manifest Anxiety Scale, the Eysenck Personality Inventory or the State-Trait Anxiety Inventory. For the studies involving children the Children’s Manifest Anxiety Scale was used.
The method of self-report can be examined first. The anxiety scales are highly structured so this suggests they can give robust and valid data. It is difficult to see how personality constructs can be measured in any other way with a large sample. However, this self-report method is not without its problems. Scales can suffer from the problem that some people will tend to choose the extreme ratings while others will cluster round the middle. Participants may not respond truthfully or may not take the questions seriously enough. It is impossible to establish cause and effect from self-reports. There may be other variables at work. There are already preconceptions in the questions so the answers are necessarily limited. Participants may answer in a way that does not accord with their views. Some participants may lack the necessary introspection to answer the questions or may not understand what they are being asked (Hoskin 2012). Thus it can be argued that anxiety scales are not entirely reliable but Hoyt and Magoon (1954) found the Taylor Manifest Anxiety Scale to be largely valid with the results according with those given by experienced counsellors. One major weakness of the study is that the self-reports may not reflect true shifts in personality but the participants’ willingness to describe themselves as anxious because of the acceptability of the term in society. Attitudes towards mental health have become less judgmental (Twenge 2000).
Secondly, the meta-analysis method will be examined. Twenge used data from many studies. This can be criticized in that the researcher has lost control of how the original studies were performed and there may have been errors in their methods. The use of meta-analysis has the advantage of enabling a large sample size to be examined suggesting the results are reliable and valid. The method avoids the bias that may be inherent in literature reviews. The method improves statistical power and improves estimates of the effect size (Noble 2006).
A difficulty for meta-analysis is the research cannot exceed the limits of what is reported by the primary research. It is a challenge to quantify the size of the common effect because of the diversity of the primary studies and their many potential differences. The effect size may be over-estimated reflecting bias in the original studies. There could still be bias in the researcher as important studies could have been left out and more favourable studies included. Meta-analysis seems very complicated so mistakes can be made which invalidate the results (Borenstein et al 2009).
In spite of some shortcomings it seems that Twenge’s study does make an important contribution to anxiety research suggesting that the larger sociocultural environment has a large impact on the mental health of the individual. There does seem a reductive element to the study though and one is left wanting to know more about the experiences of the participants. This could be achieved through a detailed case study or more open interviews.
Twenge seems to have a mix of worldviews. The major focus seems to be social constructionist. The data is qualitative from self-reports and the statistical analyses only show correlations not causations. The tone of the study seems exploratory as she is showing new possibilities for further study into the effect of society on the mental health of the individual. The participants are viewed within their contexts. There is also an element of the pragmatist worldview as this research can be used to solve real world problems: to create societies where people do not feel so anxious. There is a problem of power relations in this study as the researcher is in a more powerful position than the participant as she has framed the terms of the study. Future research on this topic could have a more transformative tone (Cresswell 2014, cited in The Open University 2017).
The second study to be examined is Brown et al (1992). This study found that anchoring events (events giving increased security) were associated with recovery or improvement in anxiety. The method used here was the semi-structured interview. The participants were interviewed twice with one year between. On the first interview information was gathered about the previous year. There was therefore clinical information for three years. The interviews were recorded and the interviewer made the final ratings. The interviewers used the DSM-III-R diagnostic system to record the anxiety symptoms. There was also a semi-structured interview for the information about life events. Raters judged the life events against what most women were likely to feel in that situation.
The semi-structured interview has many advantages as a method of psychological research. There is a pre-determined set of open questions but the interviewer can explore issues further. The participants are not limited by closed questions so the information is getting closer to their real experience that can be very valuable. There is still some uniformity because of the pre-determined questions. The structure gives reliability but there is also flexibility because of the openness. If the interviewer is skilled and well trained they may well rate the information more accurately than in the case of a self-report questionnaire. The skilled interviewer can also elicit more information about the real life events. The data can be rich (McLeod, S. A. 2014).
There are, however, obvious disadvantages. The interview is a social interaction and the participant may answer in a different way to how they would normally. The participant could be intimidated by the interviewer and tell lies or give the answers they think the interviewer wants to hear. Recording the interview could make the participant more reticent. The interviewer could suggest answers so there is a real possibility of interviewer bias. The research method is very time consuming and needs highly trained interviewers and raters (McLeod, S. A. (2014).
The use of the DSM-III-R diagnostic system can be argued to be reductive but the scale has been shown to be valid and reliable over time (Brown et al 1992). One strength of this study is that the interviewer is making the rating, not the participant, so if the interviewer is well trained the ratings should be accurate. The semi-structured interview about the anchoring and other life events can be the most challenging part of the study due to the possible bias effect of the interviewer and the complexity of rating and categorizing this material. This study does try to lessen bias effects by having a team of researchers with different people rating the information to the ones doing the interviewing.
The researchers themselves point out some limitations of the study. It is impossible to rule out bias completely. There was sometimes doubt about the exact date of the clinical change. It is possible that the clinical change as in recovery from anxiety occurred before the anchoring event and this change brought about the event rather than the reverse. The sheer complexity of the subject matter means the correlations between anchoring events and recovery from anxiety may involve other factors (Brown et al 1992).
In spite of some drawbacks of the method this study seems to provide valuable and reliable, rich evidence about the influence of social factors on anxiety that has real world implications. There is a wealth of detail in the data because of the interview method that is lacking in Twenge (2000).
Again, there seems a mix of worldviews in the design and implementation of this study. The use of the semi-structured interview provides situated information placing the participants in a social context and giving validity to their experiences. This gives the study a strong social constructionist element. There are also some pragmatic features such as the desire to solve a real problem. There is still a power inequality between the researcher and the participant with the researcher framing the terms of the research so the study is not transformative (Creswell 2014, cited in The Open University 2017).
Both methods have provided important information about the effect of the social environment on mental health and both have an exploratory tone giving pointers to further research, moving the dialogue about mental health on from beyond the individual and the family. The meta-analysis has breadth while the semi-structured interview has more depth from a smaller sample.
(Borenstein, M. Hedges, L. Higgins, J. Rothstein, H. 2009) ‘Introduction to Meta-Analysis’, John Wiley and Sons, Ltd (Online) DOI: 10.1002/9780470743386 (Accessed 24th Nov 2017)
(Brown, G. Lemyre, L. Bifulco A.1992). ‘Social Factors and Recovery from Anxiety and Depressive Disorders A test of specificity’, British Journal of Psychiatry vol. 161 pp. 44-54 (Online) DOI: 10.1192/bjp.161.1.44 (Accessed 24th Nov 2017)
(Hoskin, R. 2012) ‘The dangers of self-report’, (Online) Available at http://www.sciencebrainwaves.com/the-dangers-of-self-report/(Accessed 24th Nov 2017)
(Hoyt, D. Magoon, T. 1954) ‘A validation study of the Taylor manifest anxiety scale’,
Journal of Clinical Psychology, October 1954, Vol.10(4), pp.357-361 (Online) DOI: 10.1002/1097-4679 (Accessed 24th Nov 2017)
(McLeod, S. A. 2014) ‘The interview method’ (Online) Available at www.simplypsychology.org/interviews.html (Accessed 24th Nov 2017)
(Noble, J. 2006) ‘Meta-analysis: Methods, strengths, weaknesses, and political uses’, Journal of Laboratory and Clinical Medicine Volume 147, Issue 1, Pages 7–20 (Online) DOI: http://dx.doi.org/10.1016/j.lab.2005.08.006 (Accessed 24th Nov 2017)
The Open University (2017) DD801 Worldviews and Transformative Enquiry. 8.2 Four Worldviews on Research. Available at https://learn2.open.ac.uk/mod/oucontent/view.php?id=1153517§ion=3 (Accessed 24th Nov 2017)
(Twenge, J. 2000) ‘The Age of Anxiety? Birth Cohort Change in Anxiety and Neuroticism 1952-1993’ Journal of Personality and Social Psychology, 2000, Vol.79(6), pp.1007-1021 (Online) DOI: 10.1037/0022-35184.108.40.2067 (Accessed 24th Nov 2017)