SHOULD NUMERACY BE MEASURED OBJECTIVELY OR SUBJECTIVELY?

Purpose: Numeracy, skill handling quantitative information, is likely to affect how people use decision aids and consequently their effect on decision-making.Research linking numeracy to decision making, however, has been based primarily on objective numeracy measures. In prior research, a subjective numeracy measure, the subjective numeracy scale (SNS), has been shown to correlate moderately well with objective measures (Pearson correlation 0.6-0.70), be better accepted, be faster to complete, and provide more complete data. It is unclear, however, if these advantages outweigh the lack of evidence linking subjective numeracy directly with decision making behavior and how well the SNS correlates with additional objective numeracy measures. The suitability of using the SNS to assess numeracy skills in the context of a decision aid trial is therefore uncertain. We therefore compared numeracy assessments obtained using the SNS versus 5 objective numeracy scales: the Weller abbreviated numeracy scale, the Lipkus numeracy scale, the Berlin numeracy scale, the short version of the Numeracy Understanding in Medicine (NUMI) scale, and the 6 item version of the General Health Numeracy Test (GHNT).

Method: The study population consisted of members of an online survey panel. Each study subject completed the SNS and one objective measure. The SNS/objective scale order was reversed for one-half of each study sample. A different survey was conducted for each comparison. Data analysis consisted of correlating the results of the two measures for each study participant and comparing time to complete and participants’ willingness to answer similar questions in the future. 

Result:

Results are shown in the table. Spearman rank correlations between the SNS and the objective numeracy scales were lower than those previously reported. The SNS was substantially faster to complete than all objective scales and better accepted. There were no order effects. 

	SNS	Berlin	Ghnt	Lipkus	NUMI	Weller
number		182	113	152	124	108
Correlation with SNS		0.19	0.44	0.38	0.35	0.20
% complete in < 2 minutes	71%	27%*	15%*	22% *	24%*	9%*
% willing to do again	92%	77%*	85%*	81%*	90%*	89%*

* - p < 0.01

Conclusion: These findings suggest that additional research is needed to define the most appropriate numeracy measure to use in medical decision making research and to further characterize the differences in numeracy skill measured objectively and subjectively.