“Omitting an indicator is omitting a part of the construct.”
(p. 308)
Although formative measurement has recently received
more attention, key issues regarding its properties, advantages
and limitations are not yet well understood. As
a consequence, its use in empirical studies is still rare
(Diamantopoulos, 2008; Diamantopoulos et al., 2008). Nevertheless,
as several studies have revealed (e.g., Jarvis et al.,
2003), measurement models are often affected by misspecification.
Diamantopoulos et al. (2008) refer to this issue by
emphasizing:
“Most researchers apply scale development procedures
without even questioning their appropriateness for the
specific construct at hand [. . .]. Consequently, misspecification
commonly concerns the adoption of reflective
indicators where formative indicators (and thus index
construction approaches) would be appropriate.” (p.
1208)
In our opinion, the variable integration level of accounting
systems is of formative rather than reflective nature,
as the underlying controllers’ tasks resulting in providing
accounting information for decision-making and control
purposes are not substitutes but complements. The variable is therefore conceptualized using 17 indicators,
each describing the controllers’ (sub-)tasks with respect to
management accounting. In order to cover all core aspects
of integration, we reverted to five key controller tasks
advocated in the literature (Weber and Schäffer, 2008)
as a guideline for a comprehensive conceptualization.
Thus, for each task all relevant linkages of financial and
management accounting were identified and incorporated
into the analysis. As all 17 indicators cover different
aspects of accounting system integration, they are not
interchangeable. Also, any variation in the overall level
of integration is caused by variations in one or several of
these indicators and not the other way round.
“Omitting an indicator is omitting a part of the construct.”(p. 308)Although formative measurement has recently receivedmore attention, key issues regarding its properties, advantagesand limitations are not yet well understood. Asa consequence, its use in empirical studies is still rare(Diamantopoulos, 2008; Diamantopoulos et al., 2008). Nevertheless,as several studies have revealed (e.g., Jarvis et al.,2003), measurement models are often affected by misspecification.Diamantopoulos et al. (2008) refer to this issue byemphasizing:“Most researchers apply scale development procedureswithout even questioning their appropriateness for thespecific construct at hand [. . .]. Consequently, misspecificationcommonly concerns the adoption of reflectiveindicators where formative indicators (and thus indexconstruction approaches) would be appropriate.” (p.1208)In our opinion, the variable integration level of accountingsystems is of formative rather than reflective nature,as the underlying controllers’ tasks resulting in providingaccounting information for decision-making and controlpurposes are not substitutes but complements. The variable is therefore conceptualized using 17 indicators,each describing the controllers’ (sub-)tasks with respect tomanagement accounting. In order to cover all core aspectsof integration, we reverted to five key controller tasksadvocated in the literature (Weber and Schäffer, 2008)as a guideline for a comprehensive conceptualization.Thus, for each task all relevant linkages of financial andmanagement accounting were identified and incorporatedinto the analysis. As all 17 indicators cover differentaspects of accounting system integration, they are notinterchangeable. Also, any variation in the overall levelof integration is caused by variations in one or several ofthese indicators and not the other way round.
การแปล กรุณารอสักครู่..
