| This research study examined the effects of applying quantitative analysis to narrative data collected through a qualitative research design. The data were verbatim transcriptions of audio-taped responses to unstructured interviews concerning the development of a service organization. The data were previously analyzed using qualitative techniques and the results presented in a written report.;This study used the verbatim transcripts as the raw data for quantitative analysis techniques. The results from this analysis were then compared with the results of the analysis completed using qualitative techniques.;This study began with a content analysis of the entire sample of verbatim transcripts of interviews. Interviewer's questions, coordinating conjunctions, interjections, and prepositions were excluded from the analysis. Synonyms were combined and plurals and possessives were combined with their singular nouns.;A data matrix was constructed with sentences in the rows and unique words (the variables) in the columns. The cells contained "zero" if the variable did not occur in that row or "one" if it did occur.;A correlation matrix was constructed that contained the zero-order correlations for the 918 variables. The diagonals contained the initial estimates of the communalities for each variable. Using the criteria minimum, maximum, and average correlation, number of correlations above an absolute value of 0.13, and the communality estimates, the number of variables was ultimately reduced to 83.;An exploratory factor analysis was conducted with these 83 variables. The eigenvalues showed no clear point of departure from 1, and the scree plot showed no clear point of change in direction. A range of numbers of factors were extracted and several factor solutions were examined for simple structure and for interpretability. It was discovered that the factor solution with 13 factors retained was a satisfactory solution.;The results from the two analyses, qualitative and quantitative, show that while there are some areas of differences, there appears to be greater similarity than difference. This has been shown to be true which may raise the question of why a researcher would choose to apply both types of analyses to the same set of data. |
