There are some caveats to be taken into consideration regarding the findings of this research:
1) As shown in the statistics breakdown, 85% of the respondents are ethnic Chinese. (Ethnic Chinese are 23.7% of Malaysia’s population according to Wikipedia, but figures of Malaysian Chinese or any other Malaysian ethnicity abroad are not easily available)
2) 89% of respondents are young (between 18 to 27 years of age) and the majority of them are students (62%). As shown in my regression findings for “making a difference”, young people have a tendency to be more idealistic for Malaysia’s prospects. That could signify that their answers for the questions about current Malaysia may be generally more optimistic than if a broader age range of overseas Malaysians, with a longer time spent living and working abroad, were surveyed.
3) This survey was conducted through Google Surveys and publicized through my extended social network, word-of-mouth and Facebook. This means that to some extent, the people who responded to the survey are likely to have similar demographics and possibly similar political opinions as I do.
Nevertheless, it must be emphasized that as much as possible, these factors were accounted for in analysis, and even within a narrow demographic of Malaysians who live and/or study abroad, there are a multitude of opinions. In addition, the large sample size I had available (over 800 respondents) reduces the margin of sampling and random errors. Most professional opinion polls conducted by research centres take random samples of around 1,000 people to represent the opinions of the general population.
So, while this survey would have been best conducted randomized, due to the nature of the target population (I cannot go out to a shopping mall and survey this many people) and lack of access to statistics and contact information of Malaysians abroad, this was my best means to conduct the survey.
I used SPSS software to find ANOVA, R-Squared test and linear regression analysis to study the strength of correlation between a dependent variable and many independent variables, and how much the independent variables account for the dependent variable. Correlation does not imply causation, so I also used the comments to supplement and flesh out the meaning of figures provided by statistical analysis.
Some of the questions in the survey did not end up being used very much, particularly the form of question “How much does [x] affect your decision to stay in/leave Malaysia?”, x being a question about Malaysia like its political situation or crime rate. This is because after running regressions, I found no statistical significance between a person’s opinion of x and the strength of their decision whether to leave or stay in Malaysia. It was much better to take the independent factors of x and test them against the of strength overall desire to return to Malaysia.
To contact me for further questions regarding research method, look at contact me.