An Examination of Chi Squared as a Statistical Test in a Peer Reviewed Research Article

¶ … Chi-Square Statistical Testing in Medical Research The research undertaken by Bryant et al. (1999) was designed to assess if and how post-traumatic stress disorder (PTSD), may be prevented through the use of early intervention cognitive therapy. The research was undertaken with a sample of forty-five civilians who had survived the trauma within the preceding fourteen days and would generally be expected to develop PTSD, as assessed through the acute stress disorder interview based on DSM-IV (Bryant et al., 1999). The sample were divided into three treatment groups receiving different treatments.

The results were assessed based on the treatment received, and a post treatment assessment six months later to determine whether or not they had developed PTSD. The assessment was undertaken utilising statistical analysis, providing a method of analysis that could be tested to determine whether or not there was any significant difference. This research design indicates the research variables.

The dependent variable being tested was the outcome; whether or not the individual developed PTSD, which is measured using the clinician administered PTSD scale form 2 (15) (Bryant et al., 1999), with measurements taken immediately following treatment, and six months later. With the provision of a uniform approach to measurement of the independent variable, uniformity was gained (Cresswell, 2013).

The independent variables were the type of treatment administered, one group was subject to prolonged exposure, the second sample group experienced a combination of prolonged exposure and anxiety management, and the third group received supportive counselling. Therefore, with the need to compare frequent, traditional t-tests would be ineffective (Cresswell, 2013). The output data obtained from the research included categorical data, based on the group of the individual, as well as numerical data, based on a number of individuals within each group that went on to develop PTSD.

Therefore, a paired chi-squared test was a suitable measure to be utilised, facilitating the use of categorical and discreet numerical data (Akerlind, 2005). The results of the chi-square test were used, this provided a basis for comparing the results across all three groups against the expected outcome (Cresswell, 2013).

The results of these chi-squared test indicated that following treatment, there were more people in the supportive counselling group that fulfilled the criteria for PTSD compared to the other two groups; 20% of those in the exposure and management exactly group developed PTSD, 14% of those in the prolonged exposure group, compared to 56% of those in supportive counselling only (Bryant et al., 1999).

Interestingly, the chi-square test was also utilised in the first outcome assessment immediately following treatment, and was then utilised when the six-month data was collected, on the basis of a paired chi-squared test. This result presented the same outcome, those in the supportive counselling appeared more likely to develop PTSD compared to those in the other treatments. The group least likely to have members fulfilling the criteria of PTSD were those in the prolonged exposure group (Bryant et al., 1999).

However, it is notable that although the chi-square test was utilised, other statistical tests were also applied, including that of Yates correlation, where the results agreed with the chi-square tests (Bryant et al., 1999) Interpreting the chi-squared result should be based on the data given. Therefore, the result indicates that for patients that have suffered a recent trauma, and our risk of developing PTSD, of the three treatments tested, most effective is likely to be prolonged exposure, and the least effective is likely to be supportive counselling.

When examining the research paper, and the approach adopted, the utilisation of a chi-squared test was appropriate to the data collected. There was a reasonably sized sample, which was selected on a random basis from individuals between the ages of eighteen and sixty years, who presented with trauma even resulting from vehicle -related accidents, or nonsexual assault victims, being admitted to the PTSD unit at a hospital in Sydney, Australia, with.