Welcome 2010!!!
This year is my final year. In fact, it is the scariest year in my PhD journey. I am expected to submit my thesis by October 2010. So very soon you know.
Anyway, let me update you with my current research activities. Since the last couple of months I was so bust doing the data screening and checking for my survey data. I have divided my survey data into two main categories: the quantitative data and the qualitative data.
Alhamdulillah, I have finished my preliminary analyses on the quantitative survey data. However, I am yet to write my Results chapter in order to present the preliminary findings. I know that I am miles away from ready but I am going to enjoy every step of the way in order to really understand the various statistical analyses that I have used or intend to use.
A brief report on my preliminary analyses:
Preliminary analyses
The preliminary analyses were necessary in order to inspect the data file and to explore the nature of the variables in an attempt to address the research questions. There are several stages in this process. The first stage involved using descriptive statistics to report on the frequencies of the (a) demographics of the survey participants (such as their gender, ethnic, and religious identities, age range, highest education level, and socio-economic status (SES)); (b) education and training backgrounds related to multicultural counselling (such as their previous counsellor education programs and trainings in multicultural counselling); (c) work-related information such as the name of their jobs, the settings, the types of counselling they mostly offered to clients, and their clientele; and also to report the mean and standard deviation of the (d) overall self-reported multicultural counselling competency and its subscales (i.e., awareness, knowledge, and skills). The survey data were analysed separately first (i.e., mailed versus online survey) in order to check for similarities and differences in the patterns of results.
The second stage involved using graphs such as histograms, bars, and lines to describe and explore the data. This was necessary to explore patterns and styles of differences among several categorical variables such as gender, age, ethnic group, and others. This stage is also useful to observe patterns of results between the mailed and online survey samples.
The third stage involved checking the reliability of the instrument. ‘Reliability means that a measure (or in this case questionnaire) should consistently reflect the construct that it is measuring’ (Field, 2009, p. 673). In this reliability analysis, Cronbach’s alphas were used to compute the internal consistency of the items. Cronbach’s alphas for the total scores on multicultural counselling competency for mailed and online survey samples were checked first, followed by the alpha values on each of the total scores on multicultural awareness, knowledge, and skills. Detailed results on these analyses were reported in the previous section and they were all showed similar patterns: Cronbach’s alphas of more than .70 for the overall scores on MCC, total scores on knowledge, and skills, indicating acceptable reliability value; and Cronbach’s alpha of .50 and more for total scores on multicultural awareness.
The fourth stage involved conducting tests of significance to compare means between groups to determine whether self-reported MCCs of counsellors differ significantly by preferred survey mode (mailed versus online). This analysis was purposely conducted in order to determine whether the scores from the mailed and online survey should be combined for further analyses or treated as two separate data and analyses. An independent-samples t-test was conducted to compare the total MCC scores for mailed and online survey participants. There was no significant difference in scores for mailed (M = 113.09, SD = 10.71) and online survey participants, M = 114.40, SD = 9.58; t(508) = -1.27, p=.21 (two-tailed). The magnitude of the differences in the means (mean difference = -1.31, 95% CI: -3.35 to .72) was very small (eta squared = .003). This means that only .3 per cent of the variance in MCC is explained by survey modality. Subsequent independent-sample t-tests were conducted to compare each of the total scores for multicultural awareness, knowledge, and skills, which were the three components of MCC. Similarly, there were no significant differences in scores for mailed and online survey participants with respect to participants’ multicultural awareness, knowledge, and skills. This finding clearly indicated that scores from the mailed and online surveys can be combined for conducting further analyses to address the research aims of the present study.
The fifth stage involved conducting multiple regression (i.e., ‘categorical predictors and multiple regression’) to explore relationships among several categorical variables and MCCs (Field, 2009, p. 253).
For your information, I am still not sure about the fifth stage though...
I have to read... read and read...
Till next time.
