Developing the Ideal Grandparent Scale: Initial Validity and Reliability Tests

Abstract

Extant grandparent-grandchild research identifies numerous attributes of “ideal” grandparents. To date, however, intergenerational communication researchers have not been able to adequately operationalize grandchildren’s perceptions of their grandparents as ideal. As such, the purpose of this study was to develop the Ideal Grandparent Scale (IGS) by (a) assessing its concurrent and construct validity and (b) examining its reliability. First, the IGS items were derived deductively. Second, American grandchildren (N—171) completed the IGS along with several established instruments. The results of partial Pearson correlational analyses provided support for the hypotheses, which established initial evidence of the IGS’s concurrent and construct validity. Similarly, a split-half reliability test followed by a Feldt test and a Fisher z test provided strong support for the IGS’s reliability.

KEYWORDS:

• affection
• caring
• emotional closeness
• family identity
• ideal grandparent
• social support

Notes

1. The results of previous studies examining the effects of caregiving on grandparents’ psychological and physiological health are, at best, mixed. For instance, Guo et al. (2007) reported that grandparents who serve as caregivers for their grandchildren are psychologically and physiologically healthier than grandparents who do not serve as caregivers for their grandchildren whereas Grinstead, Leder, Jensen, and Bond (2003) concluded that grandparental caregiving does not result in negative health outcomes among grandparents. Thus, to examine if grandchildren’s perceptions of their grandparents differed as a function of whether or not a grandparent had ever served as the grandchild’s primary caregiver for an extended period of time, a multivariate analysis of covariance (MANCOVA) was conducted. In accordance with both Guo et al. (2007) and Mansson and Booth-Butterfield’s (2011) studies, several demographic variables (i.e., grandparent age, ethnicity, sex, and family lineage (maternal grandmother, maternal grandfather, paternal grandmother, or paternal grandfather) and the role of parents in the GP-GC relationships (i.e., the importance of parents in the GP-GC relationship and parental encouragement) were included as control variables. The MANCOVA did not reveal a significant model Wilk’s Λ = .93, F(7, 111) = 1.24, p = .29, η² = .07, power = .51.

2. To enhance the exactness of the Feldt test results, the exact reliability coefficients (i.e., α = .947 and α = .934) were entered into the equation. This was of particular importance given that the highest reliability was rounded up and the lowest reliability was rounded down when rounding to the nearest 100th, which magnified the discrepancy between the rounded reliability coefficients (i.e., .02; .95−.93 = .02) and the exact reliability coefficients (i.e., .013; .947−.934 = .013).

3. To enhance the exactness of the Fisher z test, the exact correlation coefficients (i.e., r = .727 and r = .675) were entered into the equation. This was done because the discrepancy between the rounded correlation coefficients (i.e., .05; .73−.67 = .05) was smaller than that of the exact correlation coefficients (i.e., .052; .727−.675 = .052).