The basic assignment is to obtain a minimum of 100 comparable texts, find themes, and analyze. To make the analysis interesting (and to get an A), you must have some comparative element in the analysis. For example, you compare boys' and girls' answers, and you predict ahead of time (in the proposal) what you think the main differences will be.
The results will be presented in class, and written up in a report. You must email me a short proposal 4 weeks before the presentation. The proposal must clearly say what the research objective is, what you expect to find (your hypotheses), and where/how you will get the data. I will then let you know if your proposed project is acceptable, at which point you can begin collecting data. You can hand in your proposal before the deadline in order to get a head start.
In most cases, the texts should be collected from respondents directly. However, in some cases I will allow you to use texts that already exist. For example, you might analyze obituaries, or personal ads, or children's books, or newspaper articles or personal web pages, etc. Also, if the texts are very short, such as personal ads, I will require you to have more of them.
Here are some sample projects:
Most projects will have a 2-stage analysis process. The first stage is the subjective part, and is the hard part. This is where you look for themes in the data using all the different methods we have covered. (See handout and other materials for week of Sept 26th) In the presentation, you will want to present some representative quotes from respondents illustrating the themes.
The second stage is the quantitative part. This is where you code each piece of text according to your main themes, and count up how many respondents mentioned elements of each theme. This will allow you to compare different groups of respondents, such as men and women. For example, you will create a spreadsheet in which in the rows are respondents, and the columns are variables. Some of the variables are codes -- 1/0 variables that indicate whether a given respondent mentioned a given theme or not. Other variables are background variables, such as gender. Then, you will do cross-tabs of the codes by gender, and compute a chi-square statistic to indicate whether the genders differ in the frequency of mention of the code.
There are some studies that don't require the 2nd stage -- but you would need to check with me first to see if yours qualifies.
You should use overhead transparencies (or a PowerPoint-type presentation using a laptop). You need to introduce your topic, explain your methods, then give the results. Most of your presentation time will be spent on discussing the themes that emerged in your text analysis. You will also present crosstabs and statistical tests comparing different groups of respondents.
If you use overhead transparencies, you must put all your names on the opening slide, and remember to give me the transparencies after your presentation, in a folder. If you are using a powerpoint presentation, all the names must be on the opening slide, and you must email me the powerpoint file.
The report summarizes the study from beginning to end, discussing the objectives, the initial expectations, the data collection, the analysis methods, and the findings. In addition, you must include a spreadsheet or wordprocessing document containing all of the texts you collected. In addition (this can be a separate file), there should be a spreadsheet containing all the codes and background variables. All of these should be turned in in electronic form (i.e., on disk or via email) within one week of the presentation.