This study investigated the outcome and dynamics of the arguments of heterosexual couples who were married or living together. Data was gathered by means of tape-recorded messages and questionnaires administered during the arguments. Couples argued about personal matters they wanted to argue about.

Outcome of arguments, conceived of as damage to the relationship and arousal of affect, was predicted to be a function of strength of the relationship, severity of argument structure, content area, and order of occurrence of the argument. Thirty couples were subjects; they had five arguments each. The above variables were manipulated and order of presentation was controlled for by a Graeco-Latin Square design.

A conceptual model of relationships was examined. A language for talking about relationships going wrong, consisting of fourteen sub- relationships, was formulated.

A conceptual model of arguments was examined. An argument has four levels: Situation descriptions, personality descriptions, damage descriptions, and threats to the relationship.

The major results and conclusions were as follows:

There was an overall difference between initial relational strength and mean post-argument relational strength.

Arguments later in the five argument session seemed to be more damaging to the relationship than arguments early in the session.

The more severe the argument structure, the more damage done to the relationship, up to a point: When the argument structure is too severe, couples appear to resist the manipulation.

There is some evidence that negative affect increases and the value of positive relational traits decreases as an argument progresses. However, it appears that arguments must be of a certain length before this escalation becomes significant. It seems that arguments continue despite the wishes of the participants.

Four orthogonal factors, which may be considered components of relationships, were found by factor analyzing the sub-relationships: Stability, Romantic Love, Pragmatic Knowledge, and Interactive Efficiency.

Manipulation of content area was not successful. This was attributed to the conceptual relatedness of the sub-relationships.

An examination of various cross-couple comparisons suggests that it is empirically and conceptually valid to distinguish facts about the relationship and facts about the people in it.

Affect aroused during an argument is correlated with loss of relational strength; neither of these two variables is correlated with initial relational strength. Mean post-argument relational strength, which is highly correlated with initial relational strength, is a strong negative predictor of loss, which offers some support for the hypothesis that stronger relationships are more resistant to damage by arguments.

It is concluded that the results have applicability to arguments between heterosexual couples in real life. It is concluded that the results may not have any applicability to other kinds of relationships. [201 pp.]