From: "Barry B. Bean" Date: Mon, 07 Oct 96 09:23:03 -0500 Subject: Research Design
I'd propose that the following experiment could substantially contribute to the discussion of the relative value of comb material to harmonica sound. Should this experiment find funding and procede, several items would need further development, but this design should serve as a suitable initial model.
1) Basic Design
The experiment will be designed as a double blind test to study the effects of comb material on the perception of harmonica tone. The experiment will be replicated a minimum of three times. The null hypothesis in this case is obvioulsly that cpomb material plays no role in the perception of harmonica sound.
2) Materials
Joe Filisko craftsman> will be commissioned to create 12 harmonicas with identical reedplates and covers. Combs would be identical in shape but differ by material: 4 wood, 4 plastic, and 4 metal. Prior to final assembly, a panel of recognized harmonica players and craftsmen will examine the reedplates and covers and confirm their identical properties. Reedplates will be randomly assigned to combs and the harmonicas would be assembled.
A questionnaire will be prepared consisting of a minimum of 20 questions designed to rate each harmonica in terms of brightness, loudness, intonation, overall tone, and subjective qualities. Questions will be 1 to 6 ratings, with a space for notes and comments at the bottom of each form.
3) Subjects
100 listening subjects and 6 playing subjects will be recruited for each replication of the study. Listening subjects will be placed into one of four categories: Harmonica players, non-harmonica musicians, non-musician audiophiles, and non-musician, non-audiophiles. Placement will be determined by answers to a brief questionnaire on listening habits and musical history. An equal number of subjects from each category will be used.
Playing subjects will be chosen at random from a pool of currently active professional harmonica players.
4) Procedure
Each player will be randomly assigned to two harmonicas. The player will play a predetermined note held for 8 seconds, a chord for an equal length of time, and a major scale in quarter notes at 60, 120, and 240. The players will play in a random sequence until each harmonica has been played twice.
Listening subjects will fill out a questionnaire following each performance, and player subjects will fill out a questionairre following eich of their own performances.
It is crucial, however, that listeners NOT be told what, if any, differences exist between players and performers. Performers should be visually obscured from the audience.
5) Analysis
Questionanairre results will be analyzed for the presence of effects of comb material. Additionally, questionnaires will be subjected to statistical analysis for the presence of effects based on listener experience. The differences in listener experience and player experience will also be compared.
6) Potential problems and solutions
Problem: Subjective player effects Solution: The random assignment of players to harmonicas and predetermined performance should control for subjective player effects. It is possible that players will be able to tell the comb material simply by picking up the harmonica and make subtle changes in their performance. I am currently unaware of a solution for this problem beyond simple instructions.
Problem: Lack of control group Solution: If time and funding permit, the entire experiment should be replicated using identical harmonicas with identical combs (including material).
Problem: Room acoustics and volume Solution: a minimum of three replications (in different rooms, with different subjects) should eliminate the effect of room acoustics. The volume issue is trickier. The experiment will ideally be conducted in a purely acoustic environment, with no amplification and no risk of secondary tonal coloring. Should this prove impossible or impractical, a simple sound system consisting of a microphone. amplifier and speaker chosen for their float response throughout the full audible range may be used. It would, of course, be essential that the same system be used in all trials.
- - B.B. Bean - Have horn. Will travel. http://www.cris.com/~Bbbean
