Laserfiche WebLink
City of Tukwila My WebLink |HelpSearch Tips |About |Sign Out
Analysis and Reporting <br />Statistical Analysis <br />We performed statistical univariate and multivariate analyses and identified significant <br />differences among respondents. This analysis was beneficial in identifying trends and levels of <br />support among respondents. HR2 analyzed the effects of the independent variables on the <br />dependent variable, demand <br />Longitudinal Method <br />For Longitudinal analysis, hypothesis derived from last two studies was tested to <br />analyze trends. Techniques such as repeated measures ANOVA, regression, and <br />linear discriminant analysis were used. These tests conducted with 2011,2012,2013, <br />2015 data and 2017 data, and multivariate analysis will be used to determine whether <br />observed changes between 2011,2012,2013,2015 data and 2017 the studies are <br />statistically significant. <br />Univariate Analysis <br />Univariate analysis consists of describing and analyzing the responses by each <br />variable and for each group or area. The responses are shown graphically in most <br />cases or in table format in cases where a large number of different responses are <br />listed. Univariates such as means, the observed variability, point estimates, standard <br />deviation, kurtosis, skewness, and shape of the distribution are explained in a clear, <br />understandable manner to aid in understanding the data and relating the findings to <br />the research goals. <br />Multivariate Analysis <br />Multivariate analysis, including multiple correlation, multiple regression, Chi -squared <br />and Analysis of Variance (ANOVA) were utilized for this project. Multiple <br />correlation analysis is concerned with the associations that exist among several <br />variables. Multiple regression analysis is concerned with the nature of the relationship <br />between those variables. Regressions show how a dependent variable changes with <br />respect to a change in independent variables. Regression models can be developed <br />that predict the values of a single dependent variable, such as purchase intent, based <br />on a set of predicting or independent variables. Models such as this can help clarify <br />the relative importance of a range of influencing factors on the purchase decision. <br />Chi -squared is concerned with measures of association between category -level <br />variables. These types of analyses will be performed to identify differences and <br />relationships between subgroups. The multivariate analysis will include: <br />32 <br />