Literacy and Life Skills for Pasifika Adults - Further Investigation: Results from the Adult Literacy and Life Skills (ALL) Survey

1. A model of prose literacy

We constructed a statistical regression model, that sought to explain variation in prose literacy of Pasifika adults in terms of variation in each of age, gender, labour force status, educational attainment, language most frequently spoken when at home, first language, place of birth and income. The model was constructed in three stages: firstly, excluding time spent in formal education and income as potential explanatory factors; secondly, including income but excluding time spent in formal education as a potential explanatory factor; thirdly, including both income and time spent in formal education.

Appendix Table 1 contains the standardised regression coefficients (and their errors) which form the model of prose literacy. The following paragraphs describe the results of the model. We look in detail at the first stage of the model and then examine changes in the subsequent stages of the model. We finish with a brief summary.

Stage 1 of the model of prose literacy

Stage 1 of the model of prose literacy included the factors age, age squared (to allow for a non-linear relationship with prose literacy), gender, labour force status (either employed or student or other), language most frequently spoken when at home, first language, and place of birth.

Of all of the factors in stage 1 of the model of prose literacy, language spoken in the home was most closely associated with English prose literacy. Pasifika adults who spoke English infrequently in the home tended to have lower prose literacy skills).

After we have accounted for all other factors in stage 1 of the model, we can also see that:

• When compared with those of middle age, both younger and older Pasifika adults tended to have lower prose literacy skills.
• Pasifika adults who spoke English as an additional language and especially, those who spoke English infrequently in the home, tended to have lower prose literacy skills.
• Pasifika men and those Pasifika adults born outside of New Zealand tended to have lower prose literacy skills.
• Pasifika adults who were employed tended to have higher prose literacy skills.
• Pasifika adults who were students tended to have higher prose literacy skills.
  

Stage 2 of the model of prose literacy

In addition to the factors included in stage 1 of the model of prose literacy, stage 2 also included income.

Of all of the factors in stage 2 of the model of prose literacy, language spoken in the home was still most closely associated with English prose literacy. Again, Pasifika adults who spoke English infrequently in the home tended to have lower prose literacy skills.

The findings for stage 1 of the model for prose literacy still held in the stage 2 model (that is, regardless of level of income) except that, after accounting for all other factors in stage 2 of the model:

• Pasifika adults with higher incomes tended to have higher prose literacy skills.
• As might be expected, those Pasifika adults who were employed tended to have higher incomes than others and income was more strongly associated with higher prose literacy skills than being employed.
  

Stage 3 of the model of prose literacy

In addition to the factors included in stage 2 of the model of prose literacy, stage 3 also included educational attainment (in the form of time spent in formal education).

Of all of the factors in stage 3 of the model of prose literacy, time spent in formal education was most closely associated with English prose literacy. Pasifika adults who had spent more time in formal education tended to have higher prose literacy skills.

The findings for stage 2 of the model for prose literacy still held in the stage 3 model (that is regardless of time spent in formal education) except that, after accounting for all other factors in stage 3 of the model:

• Those Pasifika adults who were students tended to have spent more time in formal education than others, and time in formal education was more strongly associated with higher prose literacy skills than student status.

Summary of the model of prose literacy

Of all of the factors in the model of prose literacy, time spent in formal education was most closely associated with English prose literacy. Pasifika adults who had spent more time in formal education tended to have higher prose literacy skills.

If educational attainment of Pasifika adults is not taken into consideration, language spoken in the home was most closely associated with English prose literacy. Pasifika adults who spoke English infrequently in the home tended to have lower prose literacy skills.

After we have accounted for all other factors in the model we can also see that:

• When compared with those of middle age, both younger and older Pasifika adults tended to have lower prose literacy skills.
• Pasifika adults who spoke English as an additional language and those who spoke English infrequently in the home tended to have lower prose literacy skills.
• Pasifika men and those Pasifika adults born outside of New Zealand tended to have lower prose literacy skills.
• Pasifika adults who were employed tended to have higher prose literacy skills. However, as might be expected, those who were employed tended to have higher incomes than others and income was more strongly associated with higher prose literacy skills than employment.
• Pasifika adults who were students tended to have higher prose literacy skills. However, those Pasifika adults who were students tended to have spent more time in formal education than others, and time in formal education was more strongly associated with higher prose literacy skills than student status.
• The above statements are valid regardless of the amount of time Pasifika adults spent in formal education. However, those Pasifika adults who had spent longer in formal education tended to have higher prose literacy skills.
  

Figure 1 below shows the relative impacts of the explanatory variables in the stage 3 model of prose literacy. For each explanatory variable, an estimate of the standardised regression coefficient is presented together with a 95 percent confidence interval. The confidence interval straddles zero exactly when the explanatory variable is not significantly associated with prose literacy (after controlling for the other variables in the model).

Figure 1: Standardised regression coefficients for prose literacy

The strength of association of educational attainment (“time in formal education”) and familiarity with English (“English infrequent in home”) with prose literacy is clear in Figure 1.