Chapter 3 - METHODOLOGY
Overview of Design and Research Questions
Due the nature of previous knowledge development, the research design deemed appropriate to answer the research questions is a pre-experimental, one-shot, case study design (Anastas, 1994; Cook and Campbell, 1979; DePoy, 1993). The design was selected to answer the following seven research questions:
1. What is the extent of familiarity with bluegrass music among respondents?
2. What are the tested and reported preferences for bluegrass music among respondents?
3. What is the relative reported preference for bluegrass music?
4. What association is revealed between tested and reported preferences?
5. What stereotypes are elicited among the respondents by bluegrass music?
6. What are the relationships among familiarity, reported preference and stereotypes?
7. What are the demographic correlates of bluegrass familiarity, preference, and stereotype?
The pilot study, conducted by DePoy (1995a) and discussed in Chapter 2, synthesized with the degree of theory development in the literature, has provided the foundation for the particular design selection. Previous descriptive work on familiarity, preference, and stereotype has been reported. Therefore it is appropriate that this study advance the research agenda to an examination of an association among those three variables as a basis for further developing an understanding of the position of bluegrass music in contemporary popular culture. The survey design that was selected provided the opportunity to examine the associations discussed in the research questions in a large sample of respondents in more than one location. Pre-experimental one-shot case studies allow for measurement of defined constructs (Cook and Campbell, 1979) and yield data sufficient for analysis of the level of research questions in this study.
To examine preference, tested and reported preference were measured as well as listening preference by itself. These steps were taken to address the limitation of the NEA study which only asked for reported preference and to illuminate defintional elements of preference. As discussed in detail below, this study relied on a paper-and-pencil questionnaire for data collection.
Sampling Frame, Sample and Recruitment procedures
A non-random, purposive sample of respondents was recruited for this study (Babbie, 1992). To keep the study manageable, while attempting to obtain respondents from diverse socio-economic and age groups, respondents were recruited from undergraduate college students in two states. Undergraduate students were selected because their traditional age group (17-22 year-olds) represented the largest segment of music consumers (DePoy, 1995b). These students represent a diverse group in age, background, and geographic location. The two locations chosen for data collection were Arizona and Maine.
Recruitment of respondents was accomplished by contacting faculty members who were willing to set aside twenty minutes of class time for data collection. The classes selected included biology, European history, music appreciation (for non-music majors) and public relations. The majority of the students had no specific educational interest in music. Statistics on sample characteristics are presented in Chapter 4.
A paper-and-pencil questionnaire consisting of four sections was developed for this study. To ascertain familiarity of different music genres, each respondent was asked to identify the ten types of music being played by writing in the name of the type of music heard in the blank space in Section One (See Appendix).
The ten selections of music were audio-taped and played to respondents. Each music selection for this section was chosen to accurately represent a genre. Presented below are the tunes in the order in which they were played on the audio portion of the questionnaire. Information provided includes: genre, artist, song title, and rationale for including the song in the questionnaire.
Section I, Song # 1, Classical: “Symphony No. 40” by Mozart (Mozart, 1990, track 1). Mozart represents the Classical period of vernacular classical music. Therefore, his music is described as classical by laypersons and knowledgeable “art music” listeners.
Section I, Song # 2, Big-band: “And the angels sing” by Johnny Mercer (Mercer, 1994, track 9). This performance is by Benny Goodman and his Orchestra featuring Martha Tilton’s vocal. Benny Goodman and the big-band sound are synonymous.
Section I, Song # 3, Rock: “I just want to celebrate” by Rare Earth (Rare Earth, 1993, track 6). This song has a classic rock start with a prominent rock drum and bass beat.
Section I, Song # 4, Bluegrass: “The leaves that are green” performed by The Country Gentlemen and written by Paul Simon (Simon, 1987, track 3). This recording features The Country Gentlemen, an established national touring bluegrass band with classic bluegrass instrumentation.
Section I, Song # 5, Showtunes/operettas: “Music of the night” by Andrew Lloyd Webber (Webber, 1987, track 4). The Phantom of the Opera is one of a long line of successful musicals produced by Webber for international stages and local productions. This song features the voice of the Phantom.
Section I, Song # 6, Old-time string band: “Hell among the yearlings” a traditional tune performed by Ralph Blizard & The New Southern Ramblers (Blizard, 1994, track 1). This song is included on a collection of old-time music, and as discussed in Chapter 2, represents a music style with the same roots as bluegrass but of a distinctive genre separate and apart from bluegrass.
Section I, Song # 7, Country: “Leave me something to remember you by” by Buck Owens (Owens, no date, side 2, track 2). Buck Owens’ music has helped shape and define today’s country music. This particular song starts with a strong pedal-steel guitar, the signature instrument of country music.
Section I, Song # 8, Contemporary Folk: “The kind of love you never recover from” by Christine Lavin (Lavin, 1990, track 6). Lavin’s lyrics and music are the essence of today’s singer/song writer material which is often referred to as contemporary folk music.
Section I, Song # 9, Bluegrass: “I don’t want your rambling letters” by The Stanley Brothers (Nath, no date, track 2). The Stanley Brothers are considered first generation bluegrass performers. This song represents an older style of bluegrass music popular in the 1950’s.
Section I, Song # 10, Dixieland Jazz: “Riverboat shuffle” by The Dukes of Dixieland (Carmichael, 1994, track 2). The trumpet, trombone, clarinet, piano, bass, and drums of a dixieland jazz band have a discernible style which can be easily identified from other similar styles of jazz.
The recording of these selected ten tunes was assembled on a home cassette tape recorder with a direct patch from the source machine. The narration for the instructional portion of the tape was provided by D. McLaughlin, a voice professional.
The order of the songs was chosen so that the respondent could have several successful answers before being challenged with the more difficult to identify selections. Also, the bluegrass selections and old-time selection were spaced apart from one another to diversify the selection sequence.
The selections were presented to the respondents in a classroom setting. The interviewer read aloud the instructions and operated a cassette audio player according to the questionnaire’s instructions.
To derive a bluegrass familiarity score, selections were considered as bluegrass or not bluegrass. A correct response for bluegrass familiarity was only the identification of bluegrass or not bluegrass. If a non-bluegrass tune was incorrectly identified for its genre, but was not identified as bluegrass, a correct bluegrass familiarity score was obtained. However, data were also coded as correct or incorrect for each music genre. To obtain a total familiarity score, all correct responses were summed with a possible range of 0-10, with ascending scores indicating greater familiarity.
To ascertain tested bluegrass preference, four tested preference items for each selection were included in Section One. Each of the four preference items were coded with a (1) for positive preference and a (0) for negative preference. A total bluegrass preference score, with a potential range of 0-4 for each selection was derived by summing the numeric responses coded for the preference items for the bluegrass tunes only. Preference for music listening was ascertained by the response to item #2 on each preference section. While preference scores were also derived for non-bluegrass selections, these data will not be reported because they are beyond the scope of the research questions. (Data on relative preference will be used in future comparative inquiry.) A total tested bluegrass preference score was derived by summing the selection preference totals for two bluegrass songs and dividing by 8 (the number of bluegrass preference items) to bring the range to 0-4 for ease of comparison to individual selection preference scores. This coding scheme was selected to examine the variations in preference for two diverse examples of bluegrass music. Each of the selections was also coded categorically for listening enjoyment. Including the familiarity item and tested preference items, Section One contains a total of 50 items.
To ascertain reported music preference, Section Two of the questionnaire presented each of the music genres included in the NEA study and asked respondents to rate their preference for each on a four-point scale, with ascending scores indicating increasing preference. The numeric scale was selected to provide a dispersion of scores for statistical analysis as discussed below in the section on data analysis. Similar to Section One, the results of the bluegrass preference score were analyzed for this study. The other music types were included to eliminate threats to content validity and reliability due to socially desirable response style. Data from these items will be used in future comparative study.
In Section Three of the questionnaire, stereotypes elicited by listening to bluegrass music were examined. Using the literature and the results of DePoy’s pilot study (1995a), an item pool of negative and positive stereotype statements was generated. Stereotype items were tested for content validity and the items were reduced from 10 to 8 to prevent redundancy. A four point Likert-type response set was used to ascertain agreement with each statement and to force a choice between negative and positive stereotype, while allowing the respondent a range broader than two. The items were coded from least positive (score of 1) to most positive (score of 4).
Two selections of bluegrass music, one vocal and one instrumental, were played in their entirety to the respondents:
Section 3, Song 1, Bluegrass vocal: “The Old Home Place” by J. Webb (Webb, 1995, track 1). In the liner notes this tune is stated to be a bluegrass classic. Performed by J.D. Crowe and The New South, this arrangement contains the essential elements for a bluegrass song as outlined in Chapter 2.
Section 3, Song 2, Bluegrass instrumental: “Whitewater” by Bela Fleck (Fleck, 1990, track 16). This song features nationally recognized bluegrass instrumentalists performing an improvised up-tempo tune.
The two selections were chosen to elicit stereotypes from the music sound alone for the instrumental, and from the lyric contents and sound of bluegrass vocals. After each selection, the respondent was directed to complete the items in Section Three for that selection only. Nowhere in the directions was the term bluegrass music used. This term was avoided to reduce the potential of socially desirable responses on Section Three of the questionnaire. Thus, each stereotype item was answered twice, once for each song selection.
A total stereotype score for each selection was obtained by summing the responses to the 16 stereotype items, and dividing by 16 to bring the total back to the range of 1 to 4. A stereotype score for each individual selection was also obtained by summing the eight scores for that selection and dividing by eight to return the range to 1-4. This step was taken so that comparisons among the total selection preference score, the total bluegrass preference score, and each item could be made easily. To conduct item analysis the four-point response set was collapsed into two-point categorical data, indicating favorable or non-favorable stereotypes.
Section Four was designed to obtain personal and demographic information from respondents. Gender, ethnicity, education, whether a student played an instrument or not, and family educational levels were ascertained so that group differences related to personal characteristics could be examined.
Descriptive univariate, descriptive bivariate and inferential procedures were used to analyze the data. All data were entered on SYSTAT for Windows. Frequencies were run for all items and data were cleaned before analysis was conducted.
Descriptive univariate analysis
For categorical items, frequencies and percentages were calculated and are reported in Chapter Four. For all interval level items, means and standard deviations were computed. Means and standard deviations were also computed for all indices and total scores. To answer question #1 (What is the extent of familiarity with bluegrass music among respondents?), means and standard deviations were computed for the bluegrass familiarity subtotals and total scores. To answer question #2 (What are the tested and reported preferences for bluegrass music among respondents?), means and standard deviations were computed for all tested and reported bluegrass preference subtotal and total scores. Frequencies and percentages were calculated for listening enjoyment on each bluegrass selection. Content analysis of incorrect familiarity responses was conducted to ascertain labels used to incorrectly identify bluegrass music.
To answer question #3 (What is the relative reported preference for bluegrass music?), means and standard deviations for each music genre were calculated and ranked according to mean score from most preferred to least preferred.
The Pearson correlation coefficient was calculated between tested and reported preference for bluegrass music to answer question #4 (What association is revealed between tested and reported preferences?).
To answer question #5 (What stereotypes are elicited among the respondents by bluegrass music), item analysis was completed for each stereotype item by computing frequencies, percents, means and standard deviations. For total stereotype score, means and standard deviations were computed for the total stereotype score.
To answer question #6 (What are the relationships among familiarity, reported preference and stereotypes?), a Pearson Correlation matrix was computed for total bluegrass familiarity, total reported preference, total tested preference, and total stereotype.
To answer question #7 (What are the demographic correlates of bluegrass familiarity, preference, and stereotype?), that is the relationships among preference, familiarity, age, stereotype, and educational level, a Pearson Correlation matrix was computed.
To examine group differences in preference, familiarity and stereotype related to gender, whether the respondent played an instrument or not, and geographic location, one way ANOVAs or t-tests were computed. A Post-hoc analysis using the Tukey test was computed for ANOVAs where the F ratio was significant. Group differences on stereotype scores related to listening preference were conducted by computing one-way ANOVAs with listening enjoyment as the grouping variable.
To insure that an adequate number of respondents from each of the two states could be obtained, the investigator contacted faculty members from each of the five colleges where the study was conducted and requested to administer the questionnaire in large, undergraduate survey courses. After permission was obtained, the investigator scheduled a date for data collection. Nine classes were used for data collection. At the beginning of each class session, the investigator was introduced to the students. Details were provided about the study as well as the necessary information for human subject review. Consent was noted by the completion of the questionnaire. Students were informed that they could choose whether or not to participate without penalty and that they could choose not to answer questions that made them uncomfortable. They were informed that their anonymity was preserved by not requiring identifying information on the questionnaire.
The investigator distributed the questionnaires and instructed the students to read along with the investigator’s oral instructions. As depicted in the instrumentation, each of the selections was played after which respondents were asked to complete the questions for that selection. After respondents completed Section IV, all questionnaires were collected, numbered and coded for data entry and analysis.