• Go through the spreadsheet and do any initial clean-up necessary.
  • make sure all the tokens are true /ae/-tokens
  • make all the codes consistent in each column (e.g. replace all 'F:female' with 'F')
  • check the accuracy of the coding for the linguistic variables in columns H-J and do any necessary re-coding
  • do any other clean-up/corrections you think are necessary. If you have questions about a particular token, email the person who recorded it.
  
  
• Analysis: Apply and build on the Excel tools we went over in class to investigate any questions that are of interest to you in the data. Think about the different independent variables we've talked about in class - both linguistic and nonlinguistic - and what kind of effect they might be expected to have on our dependent variable (whether /ae/ is tense or lax). Refer to part 3 of homework 1, section 3.5 of Labov 2007 (distributed in class, along with my key points summary), and your Aitchison readings for ideas about why we coded for these particular factors and what patterns we might expect to find.
  
  We went over basic sorting and formulas (Sum, SumIf, CountIf) together, but you'll probably also want to use Average, the chart wizard, and other Excel functions. The F1 button brings up the Help menu, which is usually pretty good. You can also try googling for Excel support sites, or - as always - ask a classmate, or me. If you're new to Excel, make backup copies frequently using Save As to avoid losing work in case you make a mistake.
  
  
• Write a brief report (the equivalent of 2-4 single-spaced pages, not including charts) that includes 4 to 6 main points/observations about the data. You should make it clear: (i) why you chose to investigate each variable (or combination of variables), demonstrating your understanding of the issues raised in class and in the readings cited above; (ii) whether the result was consistent with any initial expectations you had; and (iii) what you think might be some possible reasons for any surprising results. For each observation, you should include a general statement about your finding as well as specific numbers to back it up: 'Women were found to have more frequent /ae/-tensing than men (38% (n=150) vs. 20% (n=160)'. Include tables and/or charts judiciously, in cases where they help make complex findings easier to process at a glance. In many (perhaps most) cases, you might think you can't draw definite conclusions because you don't have enough data, or the right kind of data. Include these comments in your report as well.
  
  
• After reporting your results, write an additional paragraph or two (1/2 to 1 page) with a proposal for follow-up investigation. What question(s) about Philly /ae/ are you interested in pursuing and how does this initial data set need to be supplemented in order for you to pursue them? Think about the various methods of linguistic data collection discussed in class on day 1 and in the Aitchison chapters, and carefully consider which method(s) would be best suited for your questions of interest and for the time and resources we have at hand. Two additional methods you might want to consider include:
  
  
  • Semantic differentials: The investigator writes two words on cards (one or both words containing the linguistic variable under investigation) and asks the speaker to explain the difference between them. For example, you could write SAD and UNHAPPY on cards and ask 'What's the difference between these?' This method is designed to elicit a particular word/words without focusing the speaker's attention on their pronunciation (if anything, they'll probably think you're interested in the meanings of the words).
  • Reading passage: You write a short text that contains a number of short-a words you're interested in, and the speaker reads it out loud. While all reading tasks elicit a formal, careful style of speech, burying the short-a words in a larger text makes it less likely that the speaker will pay attention to their pronunciation of short-a in particular.
  
  
• Submit your report and proposal to me by email and in hardcopy by/at the beginning of class on Monday, July 16. Email me your final spreadsheet at the same time (don't try to print it). You'll be graded on (i) the accuracy of your coding and use of Excel tools, (ii) your demonstrated understanding of the relevant background issues we've read about and discussed, and (iii) the clarity and overall quality of your written presentation.

page maintained by Marjorie Pak