This paper explores the relationship between two different measures of adult literacy and numeracy. The Adult Literacy and Life Skills (ALL)Survey was an international survey of literacy, numeracy and problem solving skills undertaken in New Zealand in 2006. It provided information on the skills of the population, as well as wide range of background information on employment, qualifications and demographic characteristics. The Literacy and Numeracy for Adults Assessment Tool was developed to measure the literacy and numeracy of learners in New Zealand adult education settings. It was implemented from 2010. It provides results that are aligned with the Adult Literacy and Numeracy Progressions (Tertiary Education Commission, 2008a).
Exploring the relationship between these two measures can give us a clearer understanding of the Assessment Tool results by comparing the results with internationally referenced information.
These two measures were set up for different purposes. The ALL survey was intended for population analysis. Its purpose was to accurately measure the skill distribution of groups within the population. It was based on a representative sample of the national population and intended for research purposes and international comparison. The Assessment Tool is intended for individual testing. It is designed to provide an accurate measure of the skill of each respondent that can be used to inform teaching and learning (Yamamoto, 2013).
Both measures are based on a similar methodology, using a range of test items to infer the literacy or numeracy abilities of the respondents. They both use Item Response Theory to relate the test items to scales of literacy and numeracy ability (Darr, 2010; Statistics Canada & Organisation for Economic Cooperation and Development, 2011). This approach makes several assumptions about literacy and numeracy, namely that:
- literacy and numeracy proficiency can be quantified
- it is meaningful to say that one person has more proficiency than another person
- the amount of proficiency a person has is reasonably stable across time and place (i.e. it won’t be different for the same person a day later or in a different location, unlike measures such as happiness or cash in the bank)
- a person can demonstrate their proficiency by responding to a range of test items of varying difficulty
- the responses to the test items reflect differences in literacy and numeracy proficiency, rather than other factors, such as linguistic or cultural understanding.
These assumptions require that the measures have rigorous design and development. Particular attention is required to develop, test and scale test items to ensure that each item correctly tests an aspect of literacy or numeracy proficiency and provides valid and reliable results.
The literacy and numeracy scales derived from the ALL survey have been proven to have good correlation with education, employment and earning and health and well-being (Earle, 2009; Lawes, 2009; Smyth & Lane, 2009; Statistics Canada & Organisation for Economic Cooperation and Development, 2011). An analysis of the Assessment Tool results to 2011 found that “although groups of learners assessed using the Assessment Tool were clearly not representative of the adult population, the variations in Assessment Tool reading and general numeracy scores followed the same patterns as the total adult population according to the ALL survey 2006, in terms of age, gender, ethnicity, first language and educational participation” (Lane, 2012).