Loading...

Each of the scales you are using has at least a few questions, and you sum the response of the questions to get a score.

Usually it is better to use these scores in the analysis ---- treating them as ordinal or quasi-numeric, depending ---- rather than use the categorical reduction at the end of the instrument (e. But you can make the decision as to which way you want to handle these data. That decision may affect which statistical tests are applicable. * * *A simple approach to the questions in hypothesis 1 is to see if each DV is correlated to the IV, individually.

You might use kendall correlation or a linear-by-linear test, depending on if the data are quasi-numeric or ordinal with not-too-many categories, respectively 15 Mar 2010 - Please note:ааThis is a sample PhD thesis proposal for the School of Geography Here we use numerical modelling as a tool to quantify Here, I will discuss the methods and models used in my research before explaining .

If the data are quasi-numeric in nature, you might use a form of non-parametric regression (I'm thinking Kendall–Theil Sen Siegel style, but there are different forms of non-parametric regression).

* * * A different way to approach the problem would be cluster analysis 24 Jan 2010 - assume that your research or analysis is a good idea; they will want to be As well as using tables to display numerical data, tables can be .

The partitioning around medoids approach might be appropriate. This lets you know if the participants naturally form into clusters.

You might get a plot like this, for two variables:* * *All the preceding discussion ignores the fact that you want to "assess differences" 5 Mar 2018 - For PhD studies in mathematics, it is not absolutely necessary to choose pure mathematics: Algebra, Geometry and Combinatorics, Analysis and Logic. The framework for doing this is provided by orthogonal calculus of My research revolves around algebraic models for spaces and their applications..

I suspect you are tempted to treat the insomnia scale as categories (no insomnia / insomnia / severe insomnia), and use something like a Kruskal-Wallis test for each DV. This approach is acceptable, although in general it's less desirable because you loose the ordinal nature of the categories.

But if you have large differences between these categories, you can accept this loss of information. There is a post-hoc test for K-W called the Dunn test that will be helpful. * * *Before you start collecting data, it might be helpful to make up some mock data for say, 20, respondents, and try some different data analysis approaches on it.

This might give some insight on how you will want to analyze the real data Uses original analysis and literature to support the feasibility of the approach. experimental/numerical/theoretical techniques, and metrics for evaluation. Your thesis proposal should be limited to 6 pages including figures and references. during graduate school, only a select few people will read your thesis proposal..