|
|||
A total of approximately 1.172 million surveys were distributed to Alberta households. As opposed to a sample survey, a census was conducted to allow every household an opportunity to respond, which provided the ability to obtain a high level of confidence in the data. A total of 77,245 responses were received, representing an overall response rate of 6.6%. The results are statistically valid at the Alberta level with a confidence interval of 99.7%, plus or minus 1% level of precision. Approximately 22,500 responses were required to obtain this high level of confidence. Figures 1 and 2 provide a geographical breakdown of the response rates, and Figure 1 also provides the geographies’ respective confidence intervals. 1.2 Representation The survey was coded by ten standard geographies in Alberta, which represented major urban centers, as well as North, Central and South rural regions. The following bar chart compares the respondent distribution by geography, to the known household distribution by geography. The percent of respondents by geography closely matches the percent of surveys distributed to households by geography. This indicates limited bias and enhances the validity of the survey results through an accurate representation of the Alberta population. 1.3 Method of Reply Albertans were offered two methods in which to reply to the survey – through pre-paid postage mail-in or on the Internet. The responses consisted of 84.4% (or 65,211) mail-in replies, and 15.6% (12,034) from the Internet. Tests demonstrated that there was a statistically significant difference between the responses of these two methods of return. Mean scores from the Internet were slightly higher for the endowment fund, capital projects and debt payment options, and slightly lower for a sustainability reserve. Survey respondents were asked to rate their level of agreement with four statements using a 5-point scale. For ease of interpretation, the data from this scale was converted into an index, which spanned from 0 (indicating strongly disagree) to 100 (indicating strongly agree), utilizing progressive ratios. This allowed the ratings to be expressed using a simple 100-point index, which aids in understandability and interpretation. Figures 4 and 5 display average index scores for all respondents, as well as major geographies (Edmonton, Calgary, other major cities in Alberta (Other Urban), and rural Alberta). Figures 6 and 7 provide average index ratings for all detailed geographies. 2.2 General Overview of Findings On average, Albertans were in most agreement with the option of the Heritage Fund operating primarily as an endowment fund, followed by using a portion of it to pay the remaining debt, then using a portion of the Fund’s assets held as a reserve for sustainability, and finally, Albertans held lowest agreement with a portion of the Fund’s assets being used for capital projects option. These mean scores were found to be significantly different with statistical tests. All geographies followed this same trend, meaning that the opinions across Alberta were fairly uniform. Significant differences between geographies are outlined in Section 2.3. 2.3 Statistically Significant Differences between Geographies A significant difference means that, according to statistical tests, it has been proven that an actual difference in opinion exists between two groups, and, if the survey were to be repeated, the higher score would remain that to the lower score. If no significant difference is found between the mean scores of two groups, then it cannot be assumed, statistically, that one group has rated an item higher than another group, as this relationship has not been verified. The following are significant differences found between the major geography groups in Alberta:
Figures 8 and 9 display the overall distribution of responses for each interval in the scale. Figure 10 summarizes the percent of respondents into three groupings of agreement levels. As can be seen, there appears to be strong opinions on either side of the scale for most of the statements. Over 50% of respondents agreed (levels 4 or 5 on the scale) with the endowment fund and debt payment options. Oppositely, the majority (46.5%) of Albertans disagreed (levels 1 or 2 on the scale) with capital project funding. 4.0 Analysis of Heritage Fund (Endowment) to Other Options When looking specifically at how respondents answered Question 1, it is possible to analyze how Albertans felt about the other options. Those who strongly agreed with the endowment fund option tended to rate the other options very low, on average (ranged from 20’s to 30’s in index scores) – especially for capital projects. People who were more inclined to disagree with the endowment fund as an option, rated paying off the debt higher, on average. To obtain feedback from Albertans regarding options for the future of the Alberta Heritage Savings Trust Fund, Alberta Revenue undertook a full census of households in the province. Albertans were asked to provide their level of agreement for the following four statements:
The analyses of the results contained in this report are based on 77,245 responses received in total. Between October 28, 2002 and November 22, 2002, 70,362 responses were collected by mail or Internet. Mail-in responses received between November 22, 2002 and January 10, 2003 were also included (6,883 responses) to allow for any hold-ups in the mail system through the Christmas season. Mail-in surveys were hand-checked for batch submissions. More than two survey responses with similar markings and identical response patterns were removed. For the Internet, multiple submissions with the same ID were checked for variability (day/time, location, similar/identical response & comment patterns). Where response patterns were identical for a single ID, within a sequential timeframe, two replies were allowed. Whereas submissions that had the same ID, but had non-sequential submission times with randomness in response patterns (i.e. possibly from libraries, Internet cafés, worksites, etc.), were permitted. Out of the total responses received, 1294 deemed multiple submissions were removed. Data analysis was performed using SAS statistical software. A variety of statistical tests (including, but not limited to, ANOVA, T-Test, Chi-Square) were employed to identify significant differences. The survey was coded according to ten pre-assigned geographies in order to identify differences of opinion from different parts of the province. Household counts, for the purpose of the mail-out, were estimated according to the 2001 Census of Canada. Geography breakdowns are as follows:
A five-point equidistant scale was utilized for the survey in order to denote subtle differences in levels of agreement. A conservative approach, utilizing accepted statistical tests, was taken to verify differences in opinion between Albertans. In addition, scale scores in the survey are relative measures and should be considered as such when interpreting the results. |
|||
|
|