Jehn & Doucet Methodology


Drawn from:

Jehn, Karen A. and Lorna Doucet. 1996 Developing Categories from Interview Data: Text Analysis and Multidimensional Scaling. Part 1. Cultural Anthropology Methods Journal 8(2):15–16.

Jehn, Karen A. and Lorna Doucet. 1997 Developing Categories for Interview Data: Consequences of Different Coding and Analysis Strategies in Understanding Text. Part 2. Cultural Anthropology Methods Journal 9(1):1–7.

Ask 76 managers in sino-american joint ventures to describe two conflicts they were involved in. One was intra-cultural, the other was inter-cultural.
  • 76 managers
  • 152 stories in all
Respondents were asked standard set of probes.
Resulting texts were initially analyzed by finding all unique words from each set of stories and placing in alphabetical order
3 naive judges asked to write themselves a definition of conflict, then examine each list of words and pick out the ones they felt were related to conflict. To make it to final list, 2 judges had to pick it initially, and then convince the 3rd if needed to make it unanimous.
  • 252 words in intracultural list
  • 542 words in intercultural list
Judges categorized the words into categories (separately for each list)
  • can then count frequency of each category by summing frequencies of each word in the category
  • a total of 27 categories were found (including both lists)
The categories were converted into dimensions and judges were asked to rate each of the stories on the dimensions (i.e., to what extent did the informant describe the conflict as about work?)
Interrater reliability was measured and a final rating of each story was constructed.
Factor analysis of the ratings was conducted separately for each set of stories. 3 factors found in each. Qualitative analysis of the loadings to name factors. Intercultural stories:
  • interpersonal animosity & hostility
  • aggravation
  • volatility/outbursts

Intracultural stories:

  • hatred and animosity with volatility
  • disagreement based on values
  • <unclear>
2 judges asked to rate each story on clarity. Based on this, the 30 clearest stories from each group were identified.
Using just the clearest stories, 50 american managers in china asked to rate the similarity on a 9 point scale between pairs of stories. At they end, they were also asked to list the criteria they used to make judgments.
MDS was used to draw maps of resulting similarity matrices. These then interpreted for underlying dimensions Intercultural stories:
  • open vs resistant to change
  • situational causes vs individual traits
  • resolution potential based on trust
  • resolution potential based on patience

Intracultural stories:

  • high vs low cooperation
  • high vs low confrontation
  • problem-solve vs accept-as-is
  • resolved vs ongoing