Impact of Lead Exposure On IQ Scores Using Simple Linear Regression
Impact of Lead Exposure On IQ Scores Using Simple Linear Regression
To assess whether there is a significant association between blood lead levels and IQ scores. The
independent variable are:
Dependent Variables are the IQ scores; Full Score (iqf), Verbal IQ (iqv) and Performance IQ
(iqp).
Simple Linear Regression
Y = β0 + β 1 X 1+ β2 X 2+e
Where:
Y – IQ scores
β0- Intercept
β1 – Coefficient of X1
β2 – Coefficient of X2
Xs - independent variables
e - error term
From the results, the intercept coefficient is 106.60 which means that the performance IQ scores
is 106.60 if there is no lead in the blood. The regression coefficients in the model are not
statistically significant since P-value > 0.05.
Model 3: Impact of lead exposure on verbal IQ scores
From the results, the intercept coefficient is 87.24 which means that the verbal IQ scores is 87.24
if there is no lead found in the blood. The regression coefficient in the model are not statistically
significant since the P-value > 0.05.
In conclusion, hypothetically IQ scores should linearly decrease with increase in blood lead
volumes however it is evident that both lead blood levels in 1972 and 1973 are not statistically
significant in predicting the IQ scores and there is need to explore this relationship using other
confounding factors such as age, sex, and lead exposure group.
Segmentation Analysis
1. By Sex
To assess whether there is any effect of lead exposure among the males and female
children we consider the sex variable with two categories 1-male and 2 -female. Two
distinct subsets are created, containing male and female data separately.
From the results, the regression coefficients are not statistically significant since the P
value is more than 0.05.
2. By Age
Create two datasets that separate children under the mean age and those above the mean
age (8.93 years).
Model 3: Impact of lead exposure on IQ score among children under the age of 8
yrs.
Model 4: Impact of lead exposure on IQ score among children above the age of 8 yrs.
Age does not have any effect on lead exposure among the children .
There is no evidence suggesting that children in the control group had their IQ scores
affected due to blood lead volumes.
Model 5: Impact of lead exposure on IQ score among currently exposed group
From the model, id72 is slightly significant in predicting full IQ scores among the
currently exposed group meaning the chances of a child who was currently exposed to
lead in 1972 affecting their IQ scores was 0.24 times higher compared to a child in the
control or previously exposed groups.
Model 6 : Impact of Lead exposure among previously exposed group
There is no evidence that children who were previously exposed to lead had any impact
on their IQ scores.
In this study, we have used two different blood lead levels (1972 and 1973) to measure
impact in three IQ scores (full, verbal and performance). Using the 0.05 level of
significance we test the hypothesis.
H0: There is no association between blood lead levels and full IQ scores.
H1: There is association between blood lead levels and full IQ scores
From the output, the p-values are greater than 0.05, which implies that we fail to reject
the null hypothesis. Hence, we conclude that blood lead levels does not affect full IQ
scores, as there is no statistically significant changes in scores with change in blood lead
levels