Peligros durante el uso

In this section, we proceed with the analysis of competence formation during the transition from preschool to primary and from primary to secondary school. We try to shed some light on complementarities between basic cognitive, motor and noncognitive abilities and social and school achievement during the early life course. We assess the magnitude of preschool factors for school achievement in order to further examine the childhood skill multiplier (Heckman, 2007). In addition, we analyze whether higher secondary school track does still improve cognitive and noncognitive abilities.

2.5.1 Abilities at preschool age as predictors of social competencies at primary school age

We present findings from regression models in order to explain four measures of social com- petencies at primary school age. On the right-hand side of the regression equation, home resources, H and Y, and the level of IQ, MQ and P measured at preschool age are included. The results from OLS estimates, documented in Table 2.7, demonstrate substantial com- plementarities between abilities acquired during childhood and social competencies which a child achieves at primary school age, and several differences between the four competen- cies.

Table 2.7: The partial elasticity of preschool abilities and home resources for social competencies at school age (t = 8 years)

interest autonomy peer relations:

expert-rated self-rated

lags lags lags lags

Ht 1.44* 1.46* 0.07 0.10 0.76* 0.85* 0.27 0.30 Yt -0.00 -0.00 -0.04 -0.05 -0.00 -0.00 0.03 0.03 IQt−1 0.54* 0.49* 0.07 -0.23 0.12 0.06 0.13 0.12 M Qt−1 0.21* 0.15* 0.65* 0.44* 0.29* 0.24* 0.05 0.04 Pt−1 0.13* 0.14* -0.06 -0.09 0.21* 0.22* 0.05 0.07 Adj.R2 0.61 0.62 0.24 0.30 0.30 0.32 0.05 0.08 Observations 364 364 364 364 363 363 352 352

Note: All variables in natural logarithm; coefficients from OLS regressions including a constant; het- eroscedasticity robust standard errors; *significant at 5% level.

Source: MARS 1986-2003. Own calculations.

Cognitive, motor and noncognitive abilities at preschool age predict interests at primary school age, motor abilities predict autonomy, and motor and noncognitive abilities predict

peer relations. There are significant associations between the indicator of social com- petence, peers, and H, the M Qt−1 and Pt−1. Interests are additionally associated with

the IQt−1 acquired until the previous stage. Autonomy, measuring maturity in everyday

life, is solely linked to the past MQ. Contemporary H enhances both popularity among peers (peer relations), and the variety of actively followed interests (interests), according to expert ratings. The findings demonstrate complementarities in competence formation, indicating that higher basic cognitive, motor and noncognitive abilities acquired in child- hood help children to perform better with regard to social competencies at school age. Children from adverse home environments suffer twofold, due to poor investments in their abilities until preschool age and due to insufficient support during primary school age.

Surprisingly, however, there is no significant association with the self-rated peer relations. None of our observations are related to the child’s self-rating of social relationships and friendships (last column, Table 2.7). Findings from self-ratings differ from those of expert ratings for various reasons. This discrepancy might be caused, among other factors, by a self-protection mechanism employed by children at risk to cope with a situation in which emotional support continuously lacking. To overcome their misery, they rate their peer relationship as satisfactory. Another explanation is that children with lower levels of basic abilities are satisfied with a smaller variety in terms of their peer relationships and their interests.

2.5.2 Abilities at preschool age as predictors of grades at primary school age

Findings from linear regressions predicting school grades by preschool factors are summa- rized in Table 2.8. Grades for reading, spelling and math are evaluated in primary school at the age of 8 years, two years before ability tracking takes place in the German educa- tional system. All grade equations include the current H, the current Y and the cognitive, motor and noncognitive abilities measured at the age of 4.5 years. In an additional model, the IQ is divided into the verbal and the nonverbal abilities, V-IQ and NV-IQ, respectively.

Note that a negative coefficient in Table 2.8 implies a better grade. The estimates can be interpreted in terms of partial elasticities since the (natural) logarithm has been used for all variables. The IQ and P at preschool age are significantly related to better grades in reading and spelling as well as in math, with similar coefficients, while the MQ is not (Table 2.8). Persistence is an important predictor for later achievement in school, which

is in line with Duckworth and Seligman (2005). Interestingly, neither H nor Y is related to the grades received at age 8.

Considering two different dimensions of IQ, only the NV-IQ remains a significant predictor of better grades. Accordingly, the level of fundamental cognitive abilities, such as logical and figural reasoning and noncognitive abilities tends to be more important for predicting school achievement at the primary school level than verbal abilities. When interpreting this result, it has to be kept in mind that our sample included only children of German- speaking parents. However, if this result could be replicated in other samples, it would be of utmost importance for policies to foster human capital, since our results (see Table 2.6) point to a very short timeframe (until early childhood) for improving logical reasoning through improved socio-emotional home resources.

Table 2.8: The partial elasticity of abilities in t − 1 and home resources in t for school grades at the age of t = 8 years

reading spelling math

IQ V-/NV-IQ IQ V-/NV-IQ IQ V-/NV-IQ

Ht -0.10 -0.05 -0.64 -0.62 -0.49 -0.56 (0.43) (0.41) (0.43) (0.39) (0.42) (0.40) Yt -0.02 -0.02 -0.06 -0.07 -0.04 -0.05 (0.02) (0.06) (0.05) (0.05) (0.06) (0.06) IQt−1 -0.84* -0.60* -0.66* (0.19) (0.19) (0.19) NV-IQt−1 -0.96* -1.18* -1.11 (0.23) (0.21) (0.23) V-IQt−1 -0.26 0.16 0.19 (0.23) (0.21) (0.20) M Qt−1 -0.17 0.009 -0.21 0.001 -0.10 0.08 (0.13) (0.13) (0.12) (0.12) (0.12) (0.12) Pt−1 -0.32* -0.23 -0.29* -0.19* -0.25* -0.17 (0.10) (0.10) (0.10) (0.10) (0.10) (0.10) R2 0.21 0.25 0.21 0.28 0.17 0.22 Observations 327 327 322 322 327 327

Note: OLS regressions including a constant; heteroscedasticity robust standard errors are in parentheses; all variables in natural logarithm; *significant at 5% level.

2.5.3 Abilities at primary school age as predictors of higher-track school attendance

Findings from probit models predicting secondary school attendance are summarized in Table 2.9. All probit estimates for attending the Gymnasium include the stage-specific home resources H and Y and the cognitive, motor and noncognitive abilities. These are measured at primary school age (8 years), two years before tracking takes place. In a further specification, the total IQ is split into verbal and nonverbal abilities. In addition, all available lags of the three abilities are included in the probit equation to reduce a potential ability bias (Wooldridge, 2005).

IQ, MQ and P at primary school age are significantly related to the probability of at- tending the Gymnasium. The magnitude of P is lower compared to the IQ and higher compared to the MQ. If the verbal and the nonverbal IQ are considered separately, the NV-IQ tends to be slightly more important than the V-IQ.11Using all lags of ability (Table 2.9, column 3) reduces some of the coefficients in the probit equation without changing the conclusions. Home resources increase the probability of attending the Gymnasium. H is as important as the IQ, and, at this stage of transition from primary to secondary school, economic resources, Y, become relevant. If Y is 10% higher, the probability of attending the Gymnasium increases by 1.8%, all else being equal.

Table 2.9: Probit estimates for attending Gymnasium

IQ IQ + lags Ht 0.82* (0.37) 0.60* (0.38) Yt 0.15* (0.05) 0.18* (0.05) IQt−1 1.03* (0.15) 0.84* (0.19) M Qt−1 0.37* (0.15) 0.33* (0.16) Pt−1 0.49* (0.12) 0.38* (0.11) R2 _{0.29 0.32} Observations 357 357

Note: All variables in natural logarithm; column 3 contains all available lags in abilities, although not reported here: these lags are jointly significant; LR-tests: 86.18**, 71.35**: *significant at the 5% level; standard errors are in parentheses.

Source: MARS 1986-2003. Own calculations.

Basic cognitive, motor and noncognitive abilities are developed in the interaction with adequate socio-emotional home resources during the early life course. They predict school

11

achievement, help children to perform in terms of grades, and predict higher-track school attendance.

A numerical examination illustrates the importance of the stock of abilities and socio- emotional home resources (all values are taken from the estimation with all lags included) for higher-secondary school track. If the IQ at age 8 is 103.75 instead of 100 (i.e. is 10% higher), the average marginal probability of attending the Gymnasium would increases by 8.4%. If P at age 8 were 10% higher, the average marginal probability would increase by 3.8%. If H at age 11 were 10% higher, the average marginal probability would increase by 6%, and if Y at age 11 were to increase by 10%, the marginal increase in the prob- ability would be 1.8%. This suggests the existence of some credit market constraints at the age of 10 years. Bright children from poor households have a lower chance of entering a higher-track secondary school. The magnitude is moderate in our data. Out of 1,000 children from poorer households, 18 will be constrained in the school transition, according to our estimates. This demonstrates that cognitive and noncognitive ability ‘constraints’ from lower socio-emotional home resources at preschool age and/or from initial organic and psychosocial conditions are more important than credit constraints in the period of transition from primary to secondary school.