Types of tests:



How to Collect, Enter into the Spreadsheets, and Analyze Experimental DataExperiment1. Property Discrimination Ability of Group MembersThe tree tests described below – paired-sample, duo-trio, and triangle tests – can each be used to test whether the members of your group can discriminate the properties your group has decided to investigate.Types of tests:Entering Paired-sample trial results in a spreadsheet.Preference or a specific trait. Use trait.Choose a trait from your list (fruity). Select two samples (e.g. two bottles of wine). Have each member in the group try both samples 7 times (unlabeled, try in different orders). Each time, record which sample each group member picked out as more fruity. (Use Model Report for Paired Sample Test, located on class website under Experiment Documents)Record in 1 for successful and 0 for unsuccessful discrimination in the appropriate cell in the Excel spreadsheet. If overall all >90%, then making a real discrimination.Alternatively, look at Vierra’s “Taste thresholds” article, page 7, for table listing significance values.Duo-trioReference, two samples. Match sample to reference. (Can you discriminate between two different wines?)Triangle test3 samples, two of which match. Which one is not like the others? (Again, can you discriminate between two different wines?)Intersubjective comparison (spreadsheet) (see comments)Intersubjective comparison = n/mm=number of group members who can discriminate the propertyn=majority of reliable discriminatorsExample: Out of 6 people, 5 are reliable (>90%). 4 reliably say Sample 1 is sweeter and 1 reliably says Sample 2 is sweeter. In this case, m=5 and n=4. The intersubjective comparison is 4/5.Testing descriptive ability (matching tests)One person (or two people) write up a description each of 3 different wines. Have the rest of the group try to match the wines with the description. How many in the group correctly match to each wine?/number of people in the group (minus the writers) Example 5 people in group (two people writing descriptions). 1/5 correctly identify wine 1; 4/5 correctly identify wine 2; 2/5 correctly identify wine 3. What descriptors were the most/least helpful?Experiment 2. Evaluation by Score Card of Different Food ProductsEvaluation Spreadsheet according to a Score Cards Create Davis-like score cards (recall that the Davis scale went up to 20, but yours doesn’t have to) using your descriptors. Have those group members who have demonstrated in Experiment 1 that they can discriminate the quality in question, test some foods on the list for that quality, recording the quality points they think appropriate in the Excel spreadsheet.You can see what the average and variance is for each quality trait. See how much people disagreed on how “earthy” or “bitter” a wine was. Variance shows how much the group differed from each other. A variance of 0 means everyone rated the wine the same on that quality. Everyone thought the wine was the same level of “robustness.” High variance means that people differed greatly from each other on the wine’s robustness.Optional: test-retest. Try wines on a 2nd occasion using the evaluation score cards again. How much did people’s evaluations change/stay the same? Test statistically doing F-test. (talk to me if you’re interested in doing this)Another type of data is qualitative based out of discussions. As you’re doing the experiments and talking about the food samples and traits you’ve picked out, what type of discussions are you having? Is there consensus on what words mean?Things to RememberPractical:If you’re doing alcohol tasting especially, try to have crackers on hand to give your palate a rest. Might also be good for coffee or tea. For all other types of tasting, have water on hand to cleanse/rest your palate. Recall that (as a matter of policy) it is categorically prohibited for students to ingest alcoholic beverages as part of a class assignment.You may want to break this up into different sessions. If you’re testing 8 qualities on two samples, and you have to taste each 7 times (paired sample difference tests), that’s 8 x 2 x 7 = 112 tastings. Consider a spittoon. Make sure you’re just tasting. Recall that (as a matter of policy) it is contrary to the experimental protocols for students to ingest alcoholic beverages as part of the class assignment.Theoretical. In designing your experiments and collecting data have these issues in mind, which you will need to write the final paper.Is it possible to meaningfully evaluate and talk about food? (Use your group’s research, discussion, and experimentation.)Over all this testing, is there a difference in the way descriptive words vs. evaluative words are used/useful?Do the set of words you chose create a “semantic field”?Do the words actually describe physically measurable sensible qualities?How can these words be used to make evaluative or aesthetic judgments about food? (Perhaps you think they can’t.)Make sure you explain your conclusions and provide references or evidence to back up your claims. ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download