Multilevel Determinants of Forecasting Effectiveness: Individual, Dyadic, and System Level Predictors and Outcomes
MetadataShow full item record
With respect to content congruency (i.e., the imposition of higher level forecasts onto lower level entities), the dissertation examines the consequences of making decisions based on data from different levels of analyses (and with different geographic scopes).The results highlight the consequences of relying on higher level forecasts when a mismatch exists between organizational and national â footprintsâ . Using various economic variables to predict housing starts across levels, the analyses found disparate results for the lower level of analysis. The results also reveal great differences in the strength of the forecasting models between different levels of analysis and between different entities at the same level. Different combinations of variables contribute toward predicting the key dependent variable, housing starts, at different levels, and even between geographic markets at the same level of analysis.
The findings suggest that traditional organizational forecasting performed at the national level presents decision makers with a â hit or missâ scenario when trying to predict housing demand in the local markets. The inability to generate strong forecasts utilizing the same variables in different markets appears to be problematic. Thus, a â bottoms-upâ approach to the technical generation of forecasts is desirable Recommendations for both future research and practice are suggested.
- Doctoral Dissertations