Data users will find summary statistics easier to understand if there is some degree ofconsistency between the methods used for reporting current levels of subjective well-being and those used for reporting change over time or comparisons between groups. Thus, if current levels are described using the mean, ideally change over time should also be reported on this basis. Once again, threshold-based estimates offer both advantages and disadvantages. Ease of communication and sensitivity to changes around the threshold level come at the cost of failing to detect changes or differences elsewhere in the scale.15Although the overall information loss can in theory be managed through careful selectionof the threshold value (and potentially through multiple thresholds), it is not obviouswhere that threshold should be drawn. Selecting cut-offs according to the distribution of the data could result in setting different thresholds for different population groups and/or different countries, making comparisons impossible.
đang được dịch, vui lòng đợi..
