• At the beginning of elicitation, the interviewee was asked to identify the meaning of each element written on a card.
  • When a meaning was different from that of the researcher's, the interviewee was further asked to make clarifications, which were then recorded.
• Five cards numbered from 1 to 5 were arranged in a line forming a scale in front of the interviewee, with the smallest number ascending from the left to five on the right. The smaller number indicated the tendency towards similarity pole perception, while the larger number represented the opposite. The numbers were not intended to stand for any quantitative magnitude.
• The triad elicitation began with the shuffling of the cards. Three cards, randomly selected at a time, were put in front of the interviewee, who was then asked a question: "In what way are two of the three things indicated on the card similar in engendering a sense of teaching efficacy in you that makes them different from the third?"
• The two responses were written on two "labels". The label indicating the similar perception was put at the left end pole of the scale, while the contrast was placed at the right.
• All the elements were then rated according to the extent they belonged to either pole.
  • When an element was felt to have mixed feelings of the two poles, it was rated "3".
  • Those elements to which the construct could not be applied were excluded from the rating; they were recorded as "NA" on the grid form.
• The triad form of elicitation continued until all elements were involved in elicitation.
• After the triad process was completed, the interviewee was further asked, in front of all elements, if any elements shared a similar perception that had not been mentioned, but would distinguish them from the rest. This was referred to as the full context form elicitation. The rating and recording were performed as previously.
• Elicitation was completed when all relevant constructs had been elicited.

In some instances, the above process of elicitation was varied. An interviewee might express openly a number of constructs in one triad elicitation. The individual constructs were identified and the ratings of elements were allocated to the successive constructs. When an interviewee felt that all applicable constructs were exhausted shortly after a number of triad elicitations, he or she could be guided to the full context form of elicitation without rigidly engaging in further triad elicitation.