ROSMIRA NATASHA ROSLEE GS64988

INDIVIDUAL ASSESMENT

1) Descriptive analysis

. tab Bid3 Wtp3

| Wtp3

Bid3 | 0 1| Total

-----------+----------------------+----------

5| 2 18 | 20

10 | 3 17 | 20

15 | 1 19 | 20

20 | 11 9| 20

25 | 8 12 | 20

-----------+----------------------+----------

Total | 25 75 | 100

Rows (Bid3): Represents different bid amounts (5, 10, 15, 20, 25).
Columns (Wtp3): Shows the outcomes (0 for not accepted, 1 for accepted).
Total Row: Sum of outcomes for each bid.
Total Column: Sum of bids made at each outcome.
This table could represent bidding outcomes in an auction or decision-making scenario,
where participants bid amounts (Bid3) and the outcome (Wtp3) determines whether their
bid is accepted or not. The totals provide an overview of how many bids were made and the
distribution of outcomes across different bid amounts.

1) Inferential analysis (Probit and Logit regression)

Single Bound-CVM

- Probit regression:
. probit Wtp3 Bid3

Iteration 0: log likelihood = -56.233514

Iteration 1: log likelihood = -50.724423
Iteration 2: log likelihood = -50.668158
Iteration 3: log likelihood = -50.668157

Probit regression Number of obs = 100

LR chi2(1) = 11.13
Prob > chi2 = 0.0008
Log likelihood = -50.668157 Pseudo R2 = 0.0990

------------------------------------------------------------------------------
Wtp3 | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Bid3 | -.0677663 .021099 -3.21 0.001 -.1091196 -.026413
_cons | 1.768798 .3819869 4.63 0.000 1.020117 2.517478
------------------------------------------------------------------------------

Bid3: The coefficient for Bid3 is -0.0677663, which is statistically significant at the 0.001
level (p < 0.01). This negative coefficient indicates that as Bid3 increases, the probability of
Wtp3 being 1 decreases.
Constant (_cons): The intercept is 1.768798, also statistically significant (p < 0.01),
representing the baseline log-odds of Wtp3 when Bid3 is zero.

2) Willingness to pay estimation:

. nlcom (Wtp3:- _b[_cons]/_b[Bid3]), noheader

------------------------------------------------------------------------------
Wtp3 | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Wtp3 | 26.10143 3.585103 7.28 0.000 19.07475 33.1281
------------------------------------------------------------------------------

Log likelihood: This is the value of the log likelihood function at each iteration during the
estimation process. The goal of maximum likelihood estimation (MLE) in probit regression is
to maximize this value, which represents how well the model fits the data.

LR chi2: Likelihood Ratio chi-square statistic, which tests the overall significance of the
model. Here, LR chi2(1) indicates a chi-square test with 1 degree of freedom.

Prob > chi2: This is the p-value associated with the LR chi-square statistic. It tests the null
hypothesis that all coefficients in the model (except the intercept) are zero. A low p-value
(like 0.0008 in this case) suggests that the model as a whole is statistically significant.

Log likelihood: The final log likelihood value after convergence, which is -50.668157.

Pseudo R2: A measure of how well the model explains the variability in the outcome, similar
to R-squared in linear regression. In this case, the pseudo R-squared is 0.0990, indicating
that the model explains approximately 9.9% of the variability in Wtp3.

The coefficients (Coef.):

Bid3: This is the coefficient estimate for the predictor variable Bid3 in the probit regression
model. It represents the estimated change in the latent variable (which underlies Wtp3)
associated with a one-unit increase in Bid3. Here, -0.0677663 indicates that as Bid3
increases, the probability of Wtp3 (being 1, i.e., accepted) decreases.

_cons: This is the intercept term in the probit regression model. It represents the estimated
value of the latent variable when Bid3 is zero.

The probit regression model suggests that higher values of Bid3 are associated with a lower
probability of Wtp3 being 1 (accepted), controlling for other factors included in the model.
The model itself is statistically significant (based on the LR chi-square test), and Bid3 has a
statistically significant negative effect on the probability of Wtp3.
 3) Single Bound CVM Modified

probit WTP3 Bid Age EducationLevel Income

Iteration 0: Log likelihood = -55.107993
Iteration 1: Log likelihood = -48.46792
Iteration 2: Log likelihood = -48.40248
Iteration 3: Log likelihood = -48.402437
Iteration 4: Log likelihood = -48.402437
Probit regression Number of obs = 100
LR chi2(4) = 13.41
Prob > chi2 = 0.0094
Log likelihood = -48.402437 Pseudo R2 = 0.1217
-------------------------------------------------------------------------------
WTP3 | Coefficient Std. err. z P>|z| [95% conf. interval]
--------------+----------------------------------------------------------------
Bid | -.0855521 .0245158 -3.49 0.000 -.1336021 -.0375021
Age | -.0312424 .029844 -1.05 0.295 -.0897357 .0272508
EducationLe~l | .7421498 .6709807 1.11 0.269 -.5729483 2.057248
Income | -.0704242 .0600228 -1.17 0.241 -.1880667 .0472183
_cons | .2957128 2.492095 0.12 0.906 -4.588704 5.18013
---------------------------------------------------------------------------

Log likelihood: This is the value of the log likelihood function at each iteration during the
estimation process. The final converged log likelihood is -48.402437.

LR chi2: Likelihood Ratio chi-square statistic, which tests the overall significance of the
model. LR chi2(4) indicates a chi-square test with 4 degrees of freedom (corresponding to
the number of coefficients estimated). Here, the chi-square statistic is 13.41 with a p-value
(Prob > chi2) of 0.0094, suggesting that the model as a whole is statistically significant.

Log likelihood: The final log likelihood value after convergence is -48.402437.
Pseudo R2: This is a measure of how well the model explains the variability in the outcome,
similar to R-squared in linear regression. The pseudo R-squared here is 0.1217, indicating
that the model explains approximately 12.17% of the variability in WTP3.

Now, looking at the coefficients (Coefficient):

Bid: The coefficient estimate for Bid is -0.0855521. This suggests that as Bid increases by one
unit, the latent variable (underlying WTP3) decreases, thereby reducing the probability of
WTP3 being 1 (accepted), holding other variables constant.

Age: The coefficient estimate for Age is -0.0312424. This indicates that as Age increases by
one unit, the latent variable decreases, but this effect is not statistically significant at the
conventional level (p = 0.295).

EducationLevel: The coefficient estimate for EducationLevel is 0.7421498. This suggests that
higher levels of EducationLevel are associated with an increase in the latent variable,
thereby potentially increasing the probability of WTP3 being 1 (accepted). However, this
effect is not statistically significant at the conventional level (p = 0.269).

Income: The coefficient estimate for Income is -0.0704242. This suggests that as Income
increases by one unit, the latent variable decreases, but this effect is also not statistically
significant at the conventional level (p = 0.241).

_cons: This is the intercept term in the probit regression model. It represents the estimated
value of the latent variable when all independent variables (Bid, Age, EducationLevel,
Income) are zero. Here, the intercept is 0.2957128 with a large standard error (2.492095),
indicating substantial uncertainty around this estimate.

Bid is the only variable that appears to have a statistically significant effect on WTP3. Higher
bids decrease the probability of WTP3 being 1 (accepted).
Age, EducationLevel, and Income do not show statistically significant effects on WTP3 in this
model.

4) Willingness to pay estimation for Single Bound CVM

modified

* compute the mean values of independent variables

summ e2age, meanonly
scalar e2age_m=r(mean)
summ inc, meanonly
scalar inc_m=r(mean)

. nlcom (WTP:- (_b[_cons]+ Gender_m *_b[Gender]+ Age_m *_b[Age] + Marital_m *_b

> [Marital] + Education_m *_b[Education])/_b[Bid1]),noheader

------------------------------------------------------------------------------
 WTP1 | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
WTP | 23.3965 2.435957 9.60 0.000 18.62211 28.17089
------------------------------------------------------------------------------

Coefficient (WTP): The estimated value of WTP is 23.3965. Based on the coefficients of the
regression model used in the nlcom command, the predicted or estimated willingness to pay
(WTP) is approximately 23.3965 units.
Statistical Significance: The z-value (9.60) is associated with a very low p-value (0.000),
indicating that the estimated WTP value is statistically significant.

The nlcom command has computed the estimated WTP based on the specified coefficients
from your regression model. The estimated WTP of 23.3965 units is statistically significant,
suggesting that the factors included in your regression model (Gender, Age, Marital,
Education, and Bid1) collectively contribute to predicting WTP with this magnitude.