I have received 51 responses to my survey asking players of WoW Classic to rate various language behaviors as helpful or unhelpful. Although small, the sample size was sufficient to perform exploratory factor analysis (EFA) on the dataset. The initial results of EFA have proven intriguing and suggest that behaviors associated with positive communication are more complex than behaviors associated with negative communication.
The survey contained 10 variables, each rated on the same 5-point Likert scale ranging from “1 – Very Unhelpful” to “5 – Very Helpful.” The variables associated with positive language behaviors were morale boosting, praise for attaining goals, constructive critcism, and positive tone. Those ascribed to negative language behaviors were defeatism, blaming others, swearing and expressing anger, talking over others, and abusive tone. Descriptive statistics were obtained in SPSS and R with a high degree of agreement observed between both software packages. Mean values for each variable were associated with clearly positive language behaviors (mean > 3.00) or with clearly negative language behaviors (mean < 3.00); no variable was miscategorized.
Preliminary EFA reveals the presence of two factors associated with positive language behaviors and one factor associated with negative language beahviors. All factors were orthogonal. Both EFA and primary component analsysis suggest the instrument may be refined by removing or clarifying some variables. Specifically, expressions of anger and swearing may be vague as swearing, when used as an exclamation, is sometimes associated with joy. Nonetheless, a minimum of three variables require careful consideration in the development of future instruments: defeatism, positive tone, and swearing and expressions of anger.
The first domain to be considered in positive language behaviors in WoW classic focus on motivation: morale boosting and praise. This dimension could be considered etiquette. The second domain contained constructive criticism and planning; this dimension could be called strategy.
A single domain emerged for negative language behaviors. This contained blaming, talking over others, negative tone, and swearing and expressions of anger. This domain could be labelled rudeness.
Defeatism and positive tone were not strongly associated with any other variable, although positive tone and morale boosting showed weak signs of correlation. These items require scrutiny in the development of future instruments.
Negative correlations were observed between several variables. Interestingly, morale boosting and constructive criticism, while both positive language variables, were weakly negatively correlated. Social exchange theory might provide a framework for analyzing this relationship.
Confounding variables have been identified; issues concerning tone of voice must be addressed. Broadly-termed neutral language behaviors exist and will be explored in future studies; these may modulate other language behaviors. The development of subsequent instruments must include variables associated with humor, lore, backstory, and roleplay to name a few recently identified items.