Thursday, January 29, 2009

Week 4: Research Methods

There is a great deal of difference between qualitative and quantitative design, although there are some research methods that take a little bit from each category. Qualitative studies include ethnographies, case studies, and descriptive studies. This type of research aims to investigate and describe the process of the situation, be it an participant ethnography, or a world-wide internet survey. Qualitative researchers collect 6 sources of data: documentation, archival records, interviews/surveys, direct observation, participant observation, and physical artifacts. Quantitative research includes correlational research and experimental research, of which there are several types: true experiments, quasi-experiments, and many, many others. Quantitative research frequently uses randomly-selected sample groups and seeks to control as many variables as possible, especially in true experiments. It also poses a questions and a hypothesis as the basis of the research. Statistics come into play much more in quantitative research, including descriptive stats (mean, median, mode) and inferential stats (chi squared, t test, f test.) An important point from Morgan is that it is not useful to argue over the merits of one or the other, but that researchers should instead choose the method that best answers their particular question. Qualitative research designs are most useful in describing in-depth situations as they occur in “natural” settings -as things occur in “real-life.” Quantitative research designs are most useful in describing correlational and causal relationships between different phenomena or variables.

There is a clear difference between Validity and Reliability, although the two terms are often used incorrectly when talking about research. They are both forms of accuracy, but measure different things. Validity is the degree to which the researcher measures what (s)he claims to measure. Reliability is the external and internal consistency of the measurements. Both are needed for the research to have credibility (of the research methods), transferability (of the results to a new researcher), dependability (explaining results), and confirmability (repeatable results from a similar test). These four aspects can be analyzed to gauge the effectiveness of the research.

Statistical probability is a way of looking at research results. Total probabilities should always equal 1.0 and the results often form a bell-shaped distribution. Probability is a useful method to infer population distributions from the actual sample results. Significance is the degree of probability of the result occurring strictly from chance. In a well-controlled test, if a certain result is more than this- that is if it is statistically significant- then it can be inferred that the result is caused by the independent variable and not from random chance. Statistical significance is the difference between correlation and causation.

Thursday, January 22, 2009

Week 3 Readings

Question: Which of Kinneavy's modes best describes each reading and why?

Kinneavey's paper An Introduction to the Modes outlined four modes of discourse that offer four separate view on reality. These are: Description, which is writing that paints a static picture, Classification, which is writing that focuses on how items fit into groups, Evaluation, which is critical writing of a dynamic action from a finished, static viewpoint, and Narration, which is writing that focuses on dynamic change over time. Kinneavy stresses that no single piece of writing will completely and neatly fit into one of these categories, but the four modes of discourse offer a convenient organization of writing, and of viewpoints on reality. The four modes can also classify scholars into analysts, theorists, critics, or historians.

Garrett's The Elements of User Experience Chapter 2 is a even mix between Description and Classification. This chapter is a description of user experience in websites and classifies a website's appearance and functions into five categories. These five "planes" are the Surface, which makes up the visible appearance of a website, the Skeleton, which is the arrangement of items on a page, the Structure, which is the organization of a website's pages and functions, the Scope, which decides what pages and functions a website will include, and the Strategy, which is the reasoning and the impetus for each of the website's pages and functions. As you can see, this chapter uses Classification in it's Description of website design.

Miller's paper What's Practical about Technical Writing is primarily Evaluative, but it is only able to evaluate through its use of Description and Classification. She discusses differences in the definition of the term "practical." She uses Description to look at the traditional view of Technical Writing. She then seeks to Classify different viewpoints, people, and writing methods based upon these definitions and then Evaluates the differences.

Plato's Phaedrus is a Narration of two character's dialogue, although there are many sub-narrations and other Descriptions and Evaluations involved. The main story tells of the interaction between Socrates and Phaedrus and Socrates' opinions (Evaluation) on Lysias' speech about love. Many Descriptions are also needed from the viewpoints of both Socrates and Phaedurs. In this Narration, Socrates gives two separate speeches from somewhat opposing views about the nature of love, and in doing so discusses the power of rhetoric, dialectic, and writing. Socrate's offers a rather critical view of writing in his discussion with Phaedrus.

In Chapters 3 and 4 from Hackos and Redish's book User and Task Analysis, the authors primarily use Description to show us how to perform analysis on user's tasks and their environment. Kinneavy describes Analysis as the action that corresponds to the Descriptive mode. A writer can analyze, then describe. This seems to be what Hackos and Redish have done, and what they are asking their readers to do as well in usability research. Although when Describing a dynamic effort like these research methods, it is not completely static, as Kinneavy's definition of Description would tell you. These chapters also use Evaluation in their directions of what to do and what not to do, and Classification to describe the different methods of task and environmental analysis. They even use Narration throughout to give memorable examples of research issues.

Monday, January 12, 2009