What are the distinctive features of
panel data
compared to
cross sectional
data
Panel data observational units or people are followed over time. They allow the study of
individual units rather than the aggregative approach used by cross sectional data. Panel
data allows us to control for unobserved individual effect (omitted variables) that are
constant. Also, in panel data, you can’t treat the set of observations in each time period as
being an independent random sample which you can after pooling independent cross
sections.
Partial effect of bdrms on price
Change in price/ change in bedrooms
β0 +β1 area+β2bdrms+β3 area×bdrms+u therefore partial effect is B2bdrms +B3area
What is a P value? If the P value was 0.022, should you reject the null at 1%?
The pvalue is the minimum significance level for which the null hypothesis would be
rejected. It is obtained by assuming the test stat corresponds to the critical value (so it is
on the boundary of the rejection region) and then finding the significance level which
implies that critical value. If the pvalue=0.022 the null would NOT be rejected at the 1%
level of significance. In statistical hypothesis testing, the pvalue is the probability of
obtaining a test statistic at least as extreme as the one that was actually observed,
assuming that the null hypothesis is true
Desirable properties of a proxy variable and suggest a proxy for family income
Properties of a good proxy variable: (i) the error term in the model is uncorrelated with
all the included explanatory variables and the proxy  and (ii) the expected value of
faminc conditional on the proxy is not related to the other explanatory variables (eg.
E[faminccigs, proxy] = E[famincproxy]). Potential proxies would include family wealth,
house value, parent’s occupation or education, (or measures for low income e.g.
unemployment status, poverty indicators; or measures of high income).Requires
discussion of a variable correlated with family income.
What is meant by the sampling distribution of an estimator? What is known about
the sampling distribution of the OLS estimator under the first 4 Gauss Markov
assumptions?
The estimator (e.g. ˆβ ) used with one random sample of data (say of size n) drawn from
the population will provide an estimate (e.g. 0.078) of the underlying population
parameter (e.g. β). A new random sample (of size n) will lead to a similar, though likely
different, estimated value due to the new draw of errors in the sample. Over many
independent samples the estimator will generate a distribution over the range of possible
values for the estimate. This is the ‘sampling distribution’ of the estimator. The statistical
properties of the estimator relate to the characteristics (.e.g the mean, variance, shape) of
the sampling distribution. Under the first 4 GM assumptions, the expected value of the
OLS estimator is equal to the population parameter (i.e. OLS is unbiased). That is, the
mean of the sampling distribution is equal to the true value of the population parameter.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
This is the end of the preview.
Sign up
to
access the rest of the document.
 Three '11
 berrick
 Econometrics, Statistical hypothesis testing, Statistical significance, significance level

Click to edit the document details