TALE OF THREE FUNCTIONS
the items were ones that individuals endorsed as ways that they use AM, we would have a clean but not very useful scale. The means and standard deviations for each item and subscale are reported in Table 4. Inspection suggests that the items show adequate variability. In addition, the most highly endorsed items (on average) provide con- ceptual confirmation of the factor interpretation presented above. These include: using AM to see what caused something bad to hap- pen (Directive function), using AM to understand who one is now (Self–continuity function), using AM to strengthen friendships (Nur- turing Relationships function), and using AM to develop closer rela- tionships with others (Developing Relationships function). 1
We also examined the subscale means and describe them here us- ing the scale level labels used by the participants. The Nurturing Re- lationships subscale ranked highest (M = 4.70; SD = .85) suggesting that this function is used “often.” The means for the Self–continuity subscale and the Directive subscale both suggest that participants “occasionally” report using AM for these reasons. Participants re- port using AM for Developing Relationships least often, that is, “seldom” to “occasionally.”
VALIDATION OF THE FOUR TALE FACTORS
In this section, we examine the convergent validity of the four–factor solution by relating the subscales constructed from the results of the factor analysis to overall frequencies of thinking and talking about the past, and, more importantly, to the subscales of an established measure, the RFS (Webster, 1993, 1997).
Preliminary Convergent Validity. To begin validation of our inter- pretation of the four TALE factors, we correlated responses on each of the four subscales derived from the TALE with the overall fre- quency of thinking about and talking about one’s past. Correlations are presented in Table 5. Only correlations with alpha levels less than .008 (alpha determined by using Bonferroni correction) and effect sizes of at least .40 (representing 16% of the variance) are interpreted. As expected, the Self–Continuity subscale correlates with thinking about life and the Nurturing Relationships subscale correlates with
1. Means for the four items that did not load on any of the factors (Q2, Q8, Q12, and Q23) ranged from 3.09 (seldom) to 4.25 (occasionally).