ECON424 pre lecture quizzes

(Pre lecture 12) Which one is correct about two cointegrated I(1) time series? Check all that apply. (Two correct answers.) A) We must use some type of Dickey-Fuller test to detect the stationarity of residuals of regression of the two variables. B) Difference between them is stationary. C) A linear combination of them is stationary. D) Regression of the first difference of two conintgrated time series detects genuine long term relationship, if there is any.

A) We must use some type of Dickey-Fuller test to detect the stationarity of residuals of regression of the two variables. C) A linear combination of them is stationary.

(Pre lecture 2) Which one is correct about population model and sample model? Check all that apply. (Two correct answers.) A) We use sample model to talk about population, even though the model is not based on population B) It is almost impossible to have population model C) Sample model is useless, because it is not based on population D) Population model and sample model are the same

A) We use sample model to talk about population, even though the model is not based on population B) It is almost impossible to have population model

(Pre lecture 4) Which one is NOT correct about goodness of fit, total variations in variable of interest, and variations in model and in residuals? Check all that apply. (One correct answer.) A) Goodness of fit measures how much of total variations in the variable of interest is explained by variations in explanatory variables (model). B) Share of variations in residuals out of total variations in variable of interest defines the measure of goodness of fit. C) Total variation is always bigger than or, in some uninteresting cases, equal to variations in residuals and in model. D) Goodness of fit is a number between zero and one, including zero and one.

B) Share of variations in residuals out of total variations in variable of interest defines the measure of goodness of fit.

(Pre lecture 6) What is correct about F-test of overall validity of the model? Check all that apply. (Two correct answers.) A) A big F-statistic (bigger than critical value) means that the model has explanatory power. B) It is one sided test of R-squared being zero versus being greater than zero. C) In rare cases the value of F-statistic can be negative. D) It is a two sided test of R-squared being zero versus not being zero.

A) A big F-statistic (bigger than critical value) means that the model has explanatory power. B) It is one sided test of R-squared being zero versus being greater than zero.

(Pre lecture 5) Which one is correct about positive and negative biases and bias towards zero and bias away from zero? Check all that apply. (Two correct answers.) A) A positive bias when the true coefficient is negative, is bias towards zero. B) A positive bias when the true coefficient is positive, is bias away from zero. C) Positive bias is the same as bias towards zero and negative bias is the same as bias away from zero. D) It is impossible to have positive bias when the true coefficient is negative.

A) A positive bias when the true coefficient is negative, is bias towards zero. B) A positive bias when the true coefficient is positive, is bias away from zero.

(Pre lecture 3) In studying the factors affecting grades at your university, which one of the following CANNOT be used as possible explanatory variable? Check all that apply. (One correct answer) A) A variable that is one if the person is a student; zero if not. B) A variable that shows number of shoes the student has. C) A variable that is one if the student is female; zero if the student is male. D) A variable for the price of house of student's parents.

A) A variable that is one if the person is a student; zero if not.

(Pre lecture 11) Which one is correct about AFC and PACF? Check all that apply. (Three correct answers.) A) ACF is a series of coefficient of correlation between the variable at time t and its lagged variables. B) PACF is a direct coefficient of correlation between the variable at time t and its lagged values, after taking the effect of other factors into consideration. C) We use AFC and PACF to identify the type of the model and its order. D) The first bar in AFC is always bigger than the first bar in PAFC.

A) ACF is a series of coefficient of correlation between the variable at time t and its lagged variables. B) PACF is a direct coefficient of correlation between the variable at time t and its lagged values, after taking the effect of other factors into consideration. C) We use AFC and PACF to identify the type of the model and its order.

(Pre lecture 8) Which one is correct about using dummy variables in regressions? Check all that apply. (Two correct answers.) A) Adding multiplication of dummy variables, similar to adding dummy variables, only SHIFTS up or down the fitted lines between variable of interest and numerical variables. B) We cannot add multiplication of two dummy variables along with the two dummies in a regression because of perfect collinearity. C) If a qualitative variable has k categories, we should define k dummy variables to cover all information. D) In dealing with gender, adding a dummy for male and another dummy for female is not accepted, because of perfect collinearity.

A) Adding multiplication of dummy variables, similar to adding dummy variables, only SHIFTS up or down the fitted lines between variable of interest and numerical variables. D) In dealing with gender, adding a dummy for male and another dummy for female is not accepted, because of perfect collinearity.

(Pre lecture 2) We are interested in studying the job market conditions for teenagers in Maryland. We randomly select 100 teenagers from College Park and send them questionnaires. Which one can be our population of interest? Check all that apply. (One correct answer.) A) All teenagers in Maryland B) 100 Teenage workers in College Park C) All teenagers in College Park D) All workers in Maryland

A) All teenagers in Maryland

(Pre lecture 3) Which one is correct about unbiasedness and consistency? Check all that apply. (One correct answer.) A) An estimator can be unbiased, consistent, both, or none. B) An unbiased estimator is consistent, and vice versa. C) A consistent estimator is necessarily unbiased, but not the other way around. D) An unbiased estimator is necessarily consistent, but not the other way around.

A) An estimator can be unbiased, consistent, both, or none.

(Pre lecture 13) Which one is correct about the three R-squared measures calculated in fixed effects and random effects regressions? Check all that apply. (Two correct answers.) A) Between R-squared is based on ignoring the variations in individual values and works with average values of individuals. B) Overall R-squared is sum of within and between R-squared. C) Overall R-squared is the average of within and between R-squared. D) Within R-squared describes how much of variations in individual scores can be explained by variation in their study time, regardless of their differences with other students.

A) Between R-squared is based on ignoring the variations in individual values and works with average values of individuals. D) Within R-squared describes how much of variations in individual scores can be explained by variation in their study time, regardless of their differences with other students.

(Pre lecture 7) Which one is the most important feature of having nonlinear variables in regression? Check all that apply. (Two correct answer.) A) Change in the level of variable of interest associated with the change in the level of an explanatory variable cannot be expressed in a constant number. B) The effect of an explanatory variable decreases as the value of explanatory variable increases. C) There is always higher order polynomials that explain the variations in the variable of interest. D) The effect of explanatory variable changes by that explanatory variable or other explanatory variables.

A) Change in the level of variable of interest associated with the change in the level of an explanatory variable cannot be expressed in a constant number. D) The effect of explanatory variable changes by that explanatory variable or other explanatory variables.

(Pre lecture 2) Which one is correct about the relationship between errors and residuals? Check all that apply. (Two correct answers.) A) For those observations from the population that are in sample, there is no relationship between residual and error. B) Errors are defined for population and residuals are defined for sample. C) For those observations from the population that are in sample, residuals and errors are the same. D) For those observations from the population that are in sample, positive (or negative) residuals are necessarily accompanied by positive (or negative) error.

A) For those observations from the population that are in sample, there is no relationship between residual and error. B) Errors are defined for population and residuals are defined for sample.

(Pre lecture 11) Which one is correct about forecasting in time series data? Check all that apply. (Two correct answers.) A) Forecasting is one of the main goals of time series analysis. B) Forecasting refers to "out-of-sample" or simply future values of variable of interest. C) Forecasting in time series is the same as prediction in cross section data. D) Forecasting the future is impossible, so we rarely discuss forecasting in time series data.

A) Forecasting is one of the main goals of time series analysis. B) Forecasting refers to "out-of-sample" or simply future values of variable of interest.

(Pre lecture 3) Which one is correct about homoscedasticity and heteroscedasticity? Check all that apply. (Two correct answers.) A) Homoscedasticity means that variation of y's around their expected values are independent of values of explanatory variables. B) The fitted line in heteroscedastic error case is flatter than the fitted line in homoscedastic error. C) Homoscedasticity means that average of error term is not related to values of explanatory variables. D) Homoscedasticity is one of the assumptions of classical regression model.

A) Homoscedasticity means that variation of y's around their expected values are independent of values of explanatory variables. D) Homoscedasticity is one of the assumptions of classical regression model.

(Pre lecture 13) Which one is correct about three type of variations in panel data of students' scores? Check all that apply. (Two correct answers.) A) If instructors believe in perfect equality of scores, between variations is zero. B) Within variation is always larger than between variations. C) If every student always get the same score, within variations is zero. D) If every student always gets the same score, overall variations is zero.

A) If instructors believe in perfect equality of scores, between variations is zero. C) If every student always get the same score, within variations is zero.

(Pre lecture 6) Which one is correct about adding a variable to the model? Check all that apply. (Two correct answers.) A) If the theory we work on says a variable should be in the model, it is better to add it. B) If a variable is not highly significant, but has a test statistic of more than 1, it deserves more investigation. C) If a variable is statistically significant, it should be there, even if its significance is because of some technical issue. D) If adding a variable increases R-squared, it deserves being in the model.

A) If the theory we work on says a variable should be in the model, it is better to add it. B) If a variable is not highly significant, but has a test statistic of more than 1, it deserves more investigation.

(Pre lecture 9) Which one is correct about logit model? Check all that apply. (Three correct answers.) A) It cannot be estimated using OLS; we use Maximum Likelihood method. B) Regular R-squared is not valid for the model. C)Its predicted probability is always in the range of zero and one. D)The estimated coefficients show the effect of an extra unit of explanatory variables on probability of dependent variable being one.

A) It cannot be estimated using OLS; we use Maximum Likelihood method. B) Regular R-squared is not valid for the model. C)Its predicted probability is always in the range of zero and one.

(Pre lecture 6) Which one is correct about adjusted R-squared? Check all that apply. (Three correct answers.) A) It increases if we add an explanatory variable that has strong explanatory power. B) It increases if we add any explanatory variable. C) It is always less than R-squared. D) It punishes adding any explanatory variable, and rewards explanatory power of variables.

A) It increases if we add an explanatory variable that has strong explanatory power. C) It is always less than R-squared. D) It punishes adding any explanatory variable, and rewards explanatory power of variables.

(Pre lecture 12) Which one is correct about a Dickey-Fuller test of unit root? Check all that apply. (Two correct answers.) A) Its alternative hypothesis is stationarity of time series. B) It cannot be applied to models with deterministic time trend and/or constant term. C) It is a test of unity of coefficient of lagged variable in and AR(1) model. D) Rejection of its null hypothesis means the data has a unit root.

A) Its alternative hypothesis is stationarity of time series. C) It is a test of unity of coefficient of lagged variable in and AR(1) model.

(Pre lecture 11) Which one is correct about autoregressive of order one? Check all that apply. (Two correct answers.) A) Its partial autocorrelation function has only one spike at the first lag. B) Its autocorrelation function has only one spike at the first lag. C) The coefficient of lagged value can be positive or negative, but its absolute value is more than one. D) The effect of shocks gradually disappears over time.

A) Its partial autocorrelation function has only one spike at the first lag. D) The effect of shocks gradually disappears over time.

(Pre lecture 2) Which one is correct about OLS estimators? Check all that apply. (Two correct answers.) A) OLS is the most widely used estimation method. B) OLS is based on minimizing sum of squared residuals. C) OLS is the only estimation method in econometric. D) OLS is based on minimizing sum of residuals.

A) OLS is the most widely used estimation method. B) OLS is based on minimizing sum of squared residuals.

(Pre lecture 7) Which one is correct about interaction between two variables in regression? Check all that apply. (One correct answer.) A) Presence of an interaction terms indicates that the effect of one variable on variable of interest depends on the value of the other explanatory variable. B) Using interaction terms is only acceptable in log-level regressions. C) The coefficient of interaction term captors the relationship between two explanatory variables. D) The sign of interaction term is always multiplication of signs of the coefficient of two explanatory variables.

A) Presence of an interaction terms indicates that the effect of one variable on variable of interest depends on the value of the other explanatory variable.

(Pre lecture 13) Which one is correct about fixed effects models and random effects models? Check all that apply. (Three correct answers.) A) Random effects model assumes that individual characteristics are completely random factors. B) Fixed effects estimators are always better than random effects estimators because they account for endogeneity problem. C) Fixed effect estimators are better than random effects estimators if there is indeed a correlation between individual characteristics and idiosyncratic error terms. D) Fixed effects model assumes that there is correlation between individual characteristics and idiosyncratic error terms.

A) Random effects model assumes that individual characteristics are completely random factors. C) Fixed effect estimators are better than random effects estimators if there is indeed a correlation between individual characteristics and idiosyncratic error terms. D) Fixed effects model assumes that there is correlation between individual characteristics and idiosyncratic error terms.

(Pre lecture 1) Which one is correct about randomized experiments? Check all that apply. (Two correct answers.) A) Randomized experiments are so valuable that justifies spending huge amount of money, whenever it is possible. B) It is impossible to run a perfect randomized experiment, so randomized experiments are useless. C) It is impossible to run randomized experiments in social sciences. D) Although randomized experiments are impossible in some cases, it is always informative to think about an "ideal" experiment, whenever we study causal effect of one variable on a variable of interest.

A) Randomized experiments are so valuable that justifies spending huge amount of money, whenever it is possible. D) Although randomized experiments are impossible in some cases, it is always informative to think about an "ideal" experiment, whenever we study causal effect of one variable on a variable of interest.

(Pre lecture 10) Which one is correct about correcting issues of heteroscedasticity and autocorrelation? Check all that apply. (One correct answer.) A) Sometimes using functional forms such as logarithmic functions corrects the issue. B) Including missing explanatory variables always corrects the issue. C) They cannot be corrected if we do not know the exact mathematical form of the issue. D) They should always be corrected using WLS.

A) Sometimes using functional forms such as logarithmic functions corrects the issue.

(Pre lecture 4) Which one is correct about tests of hypothesis? Check all that apply. (Two correct answers.) A) Tests are only about parameters in population. B) The idea of hypothesis testing is to find an evidence in sample for or against a hypothesis. C) Using hypothesis testing, we can find the values of a parameter in population with certainty. D) We may test a hypothesis about a parameter in population or a statistic in sample.

A) Tests are only about parameters in population. B) The idea of hypothesis testing is to find an evidence in sample for or against a hypothesis.

(Pre lecture 14) Which one is correct about coefficients in dif-in-dif method? Check all that apply. (Three correct answers.) A) The constant term in the regression shows the average value of outcome variable for control group before treatment. B) The coefficient of time dummy variable captures the change over time for treatment group. C) The coefficient of treatment dummy captures the difference between treatment and control group before introduction of treatment. D) The coefficient of interaction between time dummy and treatment dummy is the dif-in-dif estimator and shows the causal effect of program.

A) The constant term in the regression shows the average value of outcome variable for control group before treatment. C) The coefficient of treatment dummy captures the difference between treatment and control group before introduction of treatment. D) The coefficient of interaction between time dummy and treatment dummy is the dif-in-dif estimator and shows the causal effect of program.

(Pre lecture 4) Which one is correct about type one and type two errors. Check all that apply. (Three correct answers.) A) There is a tradeoff between two types of error, reducing one is associated to increasing the other one. B) Type I error is similar to sending an innocent person to prison. C) Type one and type two errors can be reduced to zero at the same time. D) Type I error is the error we make if we reject the null hypothesis while the hypothesis is indeed correct.

A) There is a tradeoff between two types of error, reducing one is associated to increasing the other one. B) Type I error is similar to sending an innocent person to prison. D) Type I error is the error we make if we reject the null hypothesis while the hypothesis is indeed correct.

(Pre lecture 3) What is the implication of Gauss Markov theorem? Check all that apply. (Three correct answers.) A) There may be nonlinear and more precise estimators than OLS. B) There may be biased and more precise estimators than OLS. C) All linear and unbiased estimators are less precise than OLS. D) All nonlinear and biased estimators are more precise than OLS.

A) There may be nonlinear and more precise estimators than OLS. B) There may be biased and more precise estimators than OLS. C) All linear and unbiased estimators are less precise than OLS.

(Pre lecture 13) Which one is correct about panel data? Check all that apply. (One correct answer.) A) Time-invariant unobservable factors do not change over time, so we can control for them using panel data. B) Every explanatory variable in panel data models necessarily vary over time and among individuals. C) Observable time invariant factors cannot be included in panel data analysis, because they do not change over time. D) In all panel data sets the time spans, T, is smaller than the size of the sample in each period, n.

A) Time-invariant unobservable factors do not change over time, so we can control for them using panel data.

(Pre lecture 4) Which one is correct about rejection region in a test of hypothesis? Check all that apply. (Three correct answers.) A) Two sided tests, such as test of statistical significance, have rejection regions in both left and right tail of the distribution. B) Critical values define the boundaries of rejection region. C) If the test statistic falls in the rejection region, we reject the alternative hypostasis and accept the null. D) The probability of rejection region is the same as type I error.

A) Two sided tests, such as test of statistical significance, have rejection regions in both left and right tail of the distribution. B) Critical values define the boundaries of rejection region. D) The probability of rejection region is the same as type I error.

(Pre lecture 10) Which one is correct about the effect of heteroscedasticity? Check all that apply. (Three correct answers.) A) Under heteroscedasticity, variances of OLS estimators are no longer valid. B) Under heteroscedasticity, OLS estimators are no longer unbiased. C) Under heteroscedasticity, OLS estimators are no longer BLUE. D) Under heteroscedasticity, hypothesis testing may result in incorrect decision.

A) Under heteroscedasticity, variances of OLS estimators are no longer valid. C) Under heteroscedasticity, OLS estimators are no longer BLUE. D) Under heteroscedasticity, hypothesis testing may result in incorrect decision.

(Pre lecture 9) Which one of the following qualitative variables can be used as a dependent variable? Check all that apply. (Two correct answers.) A) Union membership (member/non-member) B) Race (white/non-white) C) Home ownership (owner/non-owner) D) Gender (male/female)

A) Union membership (member/non-member) C) Home ownership (owner/non-owner)

(Pre lecture 8) Which one is correct about using multiplication of dummy variables and numerical variables in regressions? Check all that apply. (Two correct answers.) A) We use interaction between dummy variables and numerical variables to analyze the different effects of numerical variables on variables of interest for different groups in dummy variable. B) The sign of interaction between a dummy and a numerical variable is the multiplication of the sings of the dummy variable and the sign of numerical variable. C) Adding multiplications of dummy variables and numerical variables is always more explanatory than adding dummy and numerical variables separately. D) It is possible that both dummy variable and the numerical variable are statistically significant, but the interaction is completely insignificant.

A) We use interaction between dummy variables and numerical variables to analyze the different effects of numerical variables on variables of interest for different groups in dummy variable. D) It is possible that both dummy variable and the numerical variable are statistically significant, but the interaction is completely insignificant.

(Pre lecture 10) Which one is correct auto correlation? Check all that apply. (Two correct answers.) A) With autocorrelated errors, regular test statistics calculated using OLS are no longer valid. B) Autocorrelation is more likely to happen in cross section data than in time series. C) With autocorrelated errors, OLS estimators are biased. D) With autocorrelated errors, OLS estimators are no longer BLUE.

A) With autocorrelated errors, regular test statistics calculated using OLS are no longer valid. D) With autocorrelated errors, OLS estimators are no longer BLUE.

(Pre lecture 14) What is the problem of counterfactual in program evaluation? Check all that apply. (Two correct answers.) A) Members of control and treatment groups have similar features that makes them inseparable. B) At any point in time, an individual is only in treatment group or control group. C) Control group is not available in most programs. D) The average difference between outcomes in control group and treatment group is not the causal effect of treatment.

B) At any point in time, an individual is only in treatment group or control group. D) The average difference between outcomes in control group and treatment group is not the causal effect of treatment.

(Pre lecture 8) Which one is correct about dummy variables? Check all that apply. (Two correct answers.) A) When a categorical variable with no order has many categories, such as states of the United States, we can assign number to each category and use it as a single numerical variable. B) Categorical variables with natural orders, such as customers' evaluation of a commodity, among categories can be converted into regular variables OR a set of dummy variables. C) Categorical variables with no natural orders, such as types of car, among categories MUST be converted into a set of dummy variables. D) Numerical variables such as income cannot be converted into a set of dummy variables.

B) Categorical variables with natural orders, such as customers' evaluation of a commodity, among categories can be converted into regular variables OR a set of dummy variables. C) Categorical variables with no natural orders, such as types of car, among categories MUST be converted into a set of dummy variables.

(Pre lecture 10) Which one is correct for tests of detecting autocorrelation? (Two correct answer.) A) If errors are autocorrelated, Durbin-Watson test statistic is 2. B) Durbin's alternative test statistic has a chi-squared distribution. C) Durbin-Watson test is a test of detecting autocorrelation with a test statistic the takes a value between zero and four. D) If errors are autocorrelated, Durbin's alternative test statistic is close to zero.

B) Durbin's alternative test statistic has a chi-squared distribution. C) Durbin-Watson test is a test of detecting autocorrelation with a test statistic the takes a value between zero and four.

(Pre lecture 12) Which one is correct about a random walk process? Check all that apply. (Two correct answers.) A) It shows explosive behavior. B) First differencing a random walk generates a stationary time series. C) Its highly consistent behavior comes from the fact that it transfers past to future without diminishing its effect. D) It has stochastic time trend, so we can de-trend it the same way we de-trend time series with deterministic time trend.

B) First differencing a random walk generates a stationary time series. C) Its highly consistent behavior comes from the fact that it transfers past to future without diminishing its effect.

(Pre lecture 6) Which one is correct about the tradeoff between bias and variance? Check all that apply. (Two correct answers.) A) If two variables are highly correlated, the inflation in variance will be negligible. B) If the variance is very inflated, we prefer to accept a small bias to avoid inflated variance. C) We should always avoid bias, even if there is a big variance. D) If the two variables are absolutely uncorrelated, we have neither bias problem nor inflated variance problem.

B) If the variance is very inflated, we prefer to accept a small bias to avoid inflated variance. D) If the two variables are absolutely uncorrelated, we have neither bias problem nor inflated variance problem.

(Pre lecture 11) Which one is correct about stationarity? Check all that apply. (Two correct answers.) A) Stationarity is the idea that past and future are structurally different. B) In stationary data, covariance between observations a constant number. C) In stationary data, average and variance of observations remain constant over time. D) Covariance stationarity is a less restrictive version of stationarity.

B) In stationary data, covariance between observations a constant number. D) Covariance stationarity is a less restrictive version of stationarity.

(Pre lecture 8) Which one is qualitative variable and should be turned into dummy variable? Check all that apply. (Three correct answers.) A) Number of children a family has. B) Income group a person is in, such as low, middle, and high income. C) Living in one of 50 states of the United States. D)Commuting-to-work methods.

B) Income group a person is in, such as low, middle, and high income. C) Living in one of 50 states of the United States. D)Commuting-to-work methods.

(Pre lecture 14) Which one is correct about issues we face in RCTs? Check all that apply. (Three correct answers.) A) Designed treatment and control groups may not be comparable. B) Individuals may react differently simply because they are in a program. C) It is usually very costly and time consuming. D) Individuals make choice and it mixes up actual control and treatment groups.

B) Individuals may react differently simply because they are in a program. C) It is usually very costly and time consuming. D) Individuals make choice and it mixes up actual control and treatment groups.

(Pre lecture 7) Which one is correct about quadratic function in regression? Check all that apply. (Three correct answers.) A) The value of variable of interest always goes up at decreasing rate. B) It always has a minimum or a maximum point, even if the value at which this maximum or minimum happen may not be realistic. C) It has curvature, showing different value of effect at different level of explanatory variable. D) It is a nonlinear function of variables.

B) It always has a minimum or a maximum point, even if the value at which this maximum or minimum happen may not be realistic. C) It has curvature, showing different value of effect at different level of explanatory variable. D) It is a nonlinear function of variables.

(Pre lecture 11) Which one is correct about time series data? Check all that apply. (Two correct answers.) A) It is a series of observations taken at one point in time. B) It is more prone to violation of independence assumption. C) It is likely that time plays the role of an independent explanatory variable. D) Prediction and forecasting is not as important in time series data as in cross section.

B) It is more prone to violation of independence assumption. C) It is likely that time plays the role of an independent explanatory variable.

(Pre lecture 14) Which one is correct about dif-in-dif method? Check all that apply. (Two correct answers.) A) Regression of outcome variable on treatment dummy provides the causal effect of treatment. B) Its accuracy hinges on equality of secular trends of control and treatment groups over time. C) The coefficient of time dummy variable captures the change over time for treatment group. D) The differences between treatment and control group before introduction of program is uses to remove the bias.

B) Its accuracy hinges on equality of secular trends of control and treatment groups over time. D) The differences between treatment and control group before introduction of program is uses to remove the bias.

(Pre lecture 14) Which one of the following methods relies heavily on observable variables to define counterfactuals? Check all that apply. (One correct answers.) A) Instrumental variables method. B) Matching methods. C) Difference in difference method. D) Regression discontinuity method.

B) Matching methods.

(Pre lecture 5) What is the benefit of multiple regression? Check all that apply. (Two correct answers.) A) Multiple regression is bigger than simple regression, bigger models are always better. B) Multiple regression allows for including more complicated functional forms of explanatory variables. C) Multiple regressions always generate unbiased results. D) Multiple regression allows for explicit control of some features, if needed.

B) Multiple regression allows for including more complicated functional forms of explanatory variables. D) Multiple regression allows for explicit control of some features, if needed.

(Pre lecture 10) Which one is the idea of tests of detecting heteroscedasticity such as Bruesch-Pagan test? (One correct answer.) A) Regress residuals on explanatory variables and test the overall validity (significance) of entire model. B) Regress squared residuals on explanatory variables and test the overall validity (significance) of entire model. C) Regress residuals on explanatory variables and test the significance of each explanatory variable. D) Regress squared residuals on explanatory variables and test the significance of each explanatory variable.

B) Regress squared residuals on explanatory variables and test the overall validity (significance) of entire model.

(Pre lecture 6) Which one is correct about testing the joint significance of more than one variable? Check all that apply. (Three correct answers.) A) It is equivalent to a set of t-tests, each for one variable. B) The restricted model can be calculated by imposing restrictions to unrestricted model. C) It compares equivalency of two models, one with restrictions and the other one without restrictions. D) Test of overall validity of model is a special case of this test in which restricted model does not have any explanatory variable.

B) The restricted model can be calculated by imposing restrictions to unrestricted model. C) It compares equivalency of two models, one with restrictions and the other one without restrictions. D) Test of overall validity of model is a special case of this test in which restricted model does not have any explanatory variable.

(Pre lecture 7) Which one is correct about a log-log function? Check all that apply. (Three correct answers.) A) The interpretation of the slope is as always, change in the level of variable of interest divided by change in the level of explanatory variable. B) The slope coefficient is unitless. C)The slope coefficient is elasticity. D)The interpretation is based on percentage change in variable of interest and percentage change in explanatory variable.

B) The slope coefficient is unitless. C)The slope coefficient is elasticity. D)The interpretation is based on percentage change in variable of interest and percentage change in explanatory variable.

(Pre lecture 1) Refer to the equation in question 1 (y=65+3.5x ) between number of cookies (x) and grade (y). Which of the following shows the correct interpretation of the intercept. Check all that apply. (One correct answer.) A) Eating cookies guarantee a minimum of 65 in the exam. B) Eating any additional cookie increases grade by 65 points. C) A student who does not eat cookies is expected to have a grade of 65. D) A student who does not eat cookies is expected to have a grade of 3.5.

C) A student who does not eat cookies is expected to have a grade of 65.

(Pre lecture 13) Which one is correct about different estimators using panel data? Check all that apply. (Two correct answers.) A) Pooled cross section estimators fully utilize valuable information panel data contains. B) Dummy variable estimators cannot be applied in cross sectional data because of limit in number of observations compared to number of variables. C) Dummy variable estimators are technically the same as fixed effect estimators. D) The results of pooled cross section are always less reliable then the results of fixed effects models, because we do not use the time related information.

C) Dummy variable estimators are technically the same as fixed effect estimators. D) The results of pooled cross section are always less reliable then the results of fixed effects models, because we do not use the time related information.

(Pre lecture 1) We are interested in measuring the effect of eating cookies on exam grades. We collect data in a sample of 150 observations (=students) on number of cookies each student has had before exam (lets say 12 hours before exam), and we have students grades. The following equation is estimated: y=65+3.5x where y represents the grades and x represents the number of cookies. Which of the following shows the correct interpretation of the relationship between number of cookies and grades. Check all that apply. (One correct answer.) A) Eating any additional cookie increases grade by 65 points. B) Eating cookies guarantees a minimum of 65 in the exam. C) Eating an additional cookie increases grade by 3.5 points. D) Eating cookies increases grade by 3.5 points.

C) Eating an additional cookie increases grade by 3.5 points.

(Pre lecture 1) Which one is correct about the experiments on mice in lab mentioned in the text? Check all that apply. (Two correct answers.) A) To save time and cost of experiments, we can give two types of medicines (two treatments) to mice and run two experiments at once. B) If the color of the two mice vary, it violates ceteris paribus, even if it is unrelated to mouse's health. C) For causal interpretation of the results, everything should be the same for control and treatment observations. D) The mouse that does not receives medicine is used as comparison observation.

C) For causal interpretation of the results, everything should be the same for control and treatment observations. D) The mouse that does not receives medicine is used as comparison observation.

(Pre lecture 5) Which one is correct about omitted variable bias? Check all that apply. (Two correct answer.) A) If the omitted variable is relevant, there is a bias. B) If the omitted variable and included variable are correlated, there is a bias. C) If the omitted variable and included variable are correlated AND the omitted variable is relevant, there is a bias. D) Random assignment of included variables cuts the relationship between omitted variable and included variable and bring the bias to zero.

C) If the omitted variable and included variable are correlated AND the omitted variable is relevant, there is a bias. D) Random assignment of included variables cuts the relationship between omitted variable and included variable and bring the bias to zero.

(Pre lecture 12) Which one is correct about error correction model. Check all that apply. (Two correct answers). A) It requires that the residuals of long term relationship be a random walk. B) It always has a significant coefficient for residual of long term regression. C) It discovers the short term relationship after adjusting for long term relationship. D) It contains the errors from long term regression to adjust for possible long term relationship.

C) It discovers the short term relationship after adjusting for long term relationship. D) It contains the errors from long term regression to adjust for possible long term relationship.

(Pre lecture 4) Which one is correct about p-value of a test? Check all that apply. (Two correct answers.) A) If it is more than the significance level, we can reject the null and claim that the variable is significant. B) It is always less than significance level. C) It is the likelihood that the test statistic is generated "just as a matter of chance". D) In the test of statistical significance, it is twice the area more extreme than the test statistic.

C) It is the likelihood that the test statistic is generated "just as a matter of chance". D) In the test of statistical significance, it is twice the area more extreme than the test statistic.

(Pre lecture 9) Which one is correct about odd-ratio? Check all that apply. (Two correct answers.) A) It is a linear function of explanatory variable. B) It is always between zero and one. C) Its log (log odd function) is a linear function of explanatory variable. D) It is probability of dependent variable being one relative to dependent variable being zero.

C) Its log (log odd function) is a linear function of explanatory variable. D) It is probability of dependent variable being one relative to dependent variable being zero.

(Pre lecture 9) Which one is correct about linear probability model? Check all that apply. (Two correct answers.) A) It always predicts a probability that is between zero and one. B) Its interpretation is based on probability of dependent variables being zero. C) Its simplicity is so appealing that many prefer to use it despite its shortcomings. D) It is estimated using OLS method.

C) Its simplicity is so appealing that many prefer to use it despite its shortcomings. D) It is estimated using OLS method.

(Pre lecture 7) Which one is correct about nonlinear functions? Check all that apply. (One correct answer.) A) Nonlinearity in variables and in parameters are both violations of regression assumptions and are not acceptable in OLS regressions. B) Linearity in variables is violation of regression assumptions, but because they happen frequently, we accept them in OLS regressions. C) Nonlinearity in variables is not violation of regression assumptions, but nonlinearity in parameters is violation of regression assumption, and is not acceptable in OLS regressions. D) Nonlinearity in variables and in parameters do not violate any assumption and are acceptable in OLS regressions.

C) Nonlinearity in variables is not violation of regression assumptions, but nonlinearity in parameters is violation of regression assumption, and is not acceptable in OLS regressions.

(Pre lecture 9) Which one is correct about pseudo R-squared and its components? Check all that apply. (Two correct answers.) A) Log likelihood of model and log likelihood of intercept, are both positive. B) Log likelihood of model is always bigger, in absolute term, than log likelihood of intercept (a model with no explanatory variables). C) Pseudo R-squared is always between zero and one. D) If the model has no explanatory power, pseudo R-squared is zero.

C) Pseudo R-squared is always between zero and one. D) If the model has no explanatory power, pseudo R-squared is zero.

(Pre lecture 12) Which one is correct about relationship between two non-stationary I(1) time series? Check all that apply. (One correct answer.) A) If they are cointegrated, they have a long term relationship, but not a short term relationship. B) Their stochastic time trend is the only source of their correlation, and after differencing, they do not have any relationship. C) They may or may not have long term and/or short term relationship. D) They necessarily are correlated in long term.

C) They may or may not have long term and/or short term relationship.

(Pre lecture 5) Which one is correct about including an irrelevant variable (a variable whose coefficient in population is zero) in the model? Check all that apply. (One correct answer.) A) It does not matter because what we get when we estimate the model using sample will be exactly zero. B) It does matter because it tremendously increases R-squared, and give a wrong signal about the value of the model. C) It does matter, because it creates bias in estimates of the other coefficients. D) It does not matter, because based on feature 1, its expected value is zero, and even if we include it, the estimates coefficient is most likely insignificant.

D) It does not matter, because based on feature 1, its expected value is zero, and even if we include it, the estimates coefficient is most likely insignificant.

(Pre lecture 1) Why RAND experiment is considered as a very important study? Check all that apply. (One correct answer.) A) It is an example of a randomized experiment in social science in which perfect random assignment is very easy, and there is no contamination in experiment. B) It is an example of randomized experiment, and any randomized experience is important. C) It is about health and everything about health is important. D) It is an example of randomized experiment in an area where randomized experiment is almost impossible, both because of the nature of the issue and its cost.

D) It is an example of randomized experiment in an area where randomized experiment is almost impossible, both because of the nature of the issue and its cost.

(Pre lecture 8) Which one is correct about using dummy variables in regressions? Check all that apply. (One correct answer.) A) We cannot use only dummy variables as explanatory variables. There should be at least one numerical variable in the model. B) We can use as many dummy variables as we want, because unlike numerical variables, they don't reduce degrees of freedom. C) Adding unrelated dummy variables to the model reduces R-squared, so it is not useful. D) Number of dummy variables we use in a regression is limited only by the number of observations, i.e., degrees of freedom.

D) Number of dummy variables we use in a regression is limited only by the number of observations, i.e., degrees of freedom.

(Pre lecture 2) Which one of the following is NOT correct about algebraic characteristics of OLS estimators? Check all that apply. (One correct answer.) A) Sum of all residuals is zero. B) The regression line always passes through the average of the sample C) The sample covariance between x and residuals is zero. D) The regression line always passes through the origin (x=0, y=0)

D) The regression line always passes through the origin (x=0, y=0)

(Pre lecture 5) Which one is the meaning of perfect collinearity? Check all that apply. (One correct answer.) A) Two or more explanatory variables have any types of relations, linear or non-linear. B) ‌Two or more explanatory variables have correlations. C) Two or more explanatory variables have positive correlations. D) Two or more explanatory variables are perfectly related in a linear way.

D) Two or more explanatory variables are perfectly related in a linear way.

(Pre lecture 3) Which one violates the assumption of linearity of parameters? Check all that apply. (One correct answer.) A) Using squared value of an explanatory variable. B) Using a parameters as constant value in the regression. C) Using inverse of an explanatory variables. D) Using multiplication of two parameters.

D) Using multiplication of two parameters.