Goal Variability and Message Content During Relational Discussions

Abstract

Discussions about relationship issues are rarely easy, and as in any important conversation, the actions of a relational partner may encourage one to change initial message intentions and behaviors during the course of an interaction. The interrelationship of communication goals and message content was examined through a sequential analysis of 10-minute conversations between 100 dating partners about a relationship problem. After completing survey measures, respondents engaged in a discussion about one partner's problematic behavior. Respondents then reviewed discussions to rate the importance of their communication goals (self-oriented, other-oriented, relational, or task) at 1-minute intervals. Objective coders assessed the grammatical focus of respondents' messages (self, relationship, partner, or task) at the same or adjoining interval. Lag sequential analyses indicated that goal importance defined self-interested or more prorelational content, providing evidence that goals are fluid and communicative behavior is influenced by dyadic interaction.

Keywords:

• Close Relationships
• Goals
• Message Production
• Sequential Analysis

Notes

Note. Cell entries reflect the frequency of a goal category as rated as the most important goal compared to all others in a given 1-minute increment.

Note. Cell entries reflect values associated with the goal at one interval–goal at subsequent interval sequence for a given conversational interactant (initiator or partner). Entries in parentheses are conditional probabilities. Entries not in parentheses are Alison-Liker z scores, reflecting the testing of conditional against unconditional probabilities.

*p< .05. **p < .01.

Note. Cell entries reflect values associated with the goal-message-focus sequence for a given conversational interactant (initiator or partner). Entries in parentheses are conditional probabilities. Entries not in parentheses are Alison-Liker z scores, reflecting the testing of conditional against unconditional probabilities.

*p< .05. **p< .01.

Note. Entries in parentheses are conditional probabilities. Entries not in parentheses are Alison-Liker z scores, reflecting the testing of conditional against unconditional probabilities.

*p< .05. **p < .01.

Note. Entries in parentheses are conditional probabilities. Entries not in parentheses are Alison-Liker z scores, reflecting the testing of conditional against unconditional probabilities.

*p < .05. **p < .01.

The use of lag sequential analysis is contingent on the assumptions that the data are ordered, stationary, and homogeneous. Anderson Goodman likelihood ratio χ² tests (LRχ²; Anderson & Goodman, 1957) indicated that the data were first order, suggesting that an examination across increments was best captured between a 1-minute increment and the adjacent 1-minute increment. LRχ² tests indicated that the data were stationary and homogeneous at all levels across groups (i.e., initiators versus partners and males versus females), thus analyses were collapsed across groups when relevant.