Econometrics Assignment
econgirly246
econometrics_assignment.docx
a) Based on the data given in the EIVEWS file, write a suitable regression model that would allow you to investigate the correlates of attitudes towards abortion in Trinidad and Tobago.
b) State (and explain) your a-priori expectations of the coefficients in your model (Do not plagiarise. Be sure to reference!) Note, stating “we expect the coefficient on the XYZ variable to be positive/negative” is not enough.
c) Estimate your model in EVIEWS. Save your regression output in the EVIEWS file
d) Based on your results, complete the following table:
|
Variables |
Coefficients |
Standard error |
|
Diagnostics |
||
|
R-squared
e) Based on your estimated results, discuss the partial effects of your independent variables.
f) In EVIEWS, use the “partialling” out method to estimate the coefficient on the “premarital” variable. Be sure to save this regression output in the EVIEWS file. How does it compare to the coefficient presented in Table 2?
g) A researcher is only interested in the impact of age on abortion attitudes, and so, believes that he does not need to estimate the multiple regression model from (a). He only needs to estimate the following simple regression
𝑎𝑏𝑜𝑟𝑡𝑖𝑜𝑛= 𝛽0+𝛽1𝑎𝑔𝑒+𝑢
Do you agree? Carefully explain your logic (Hint: are any assumptions violated?). Provide statistical evidence in support of your argument.
h) A researcher believes that religiosity and education are the only factors that influence abortion attitudes. i. Explain how would you go about testing this hypothesis ii. Conduct the test in EVIEWS (and be sure to freeze the results). What is your decision? What is your conclusion?
i) Give one possible reason why this model could present omitted variable bias. Be sure to explain your logic.
QUESTION 3
a) Consider the following regression, which includes a constant: i. The RSS ii. The 3 × 1 matrix (the “ ^ ” is supposed to be above beta, representing estimate of ) iii. The variance-covariance matrix of (the “ ^ ” is supposed to be above beta, representing estimate of ) iv. Conduct the following hypothesis tests
𝑌 = 𝑋𝛽 + 𝑢 where 𝑌 is an 𝑛 × 1 matrix, 𝑋 is an 𝑛 × 3 matrix, 𝛽 is a 3 × 1 matrix and 𝑢 is an 𝑛 × 1 matrix. The regression was estimated using 103 observations and the following were calculated:
where 𝑠2 is the estimate of 𝑣𝑎𝑟(𝑢) = 𝜎2 Use this information to calculate the following 𝐻0: 𝛽1 = 0; 𝐻1: 𝛽1 ≠ 0 𝐻0: 𝛽1 + 𝛽2 ≥ 1; 𝐻1: 𝛽1 + 𝛽2 < 1 |
||
|
|
