PAC Research Methodologies & Data Analysis 5

Question 1

A policy analyst conducts a logistic regression to predict whether a household will fall into poverty (yes/no). Why is logistic regression preferred over ordinary least squares (OLS) in this case?

Accepted Answer

OLS can produce predicted probabilities outside the 0–1 range for binary outcomes

Answer

OLS can generate predicted values below 0 or above 1 for binary outcomes, while logistic regression constrains predictions to the valid probability range [0,1].

Question 2

In policy research, 'triangulation' refers to:

Accepted Answer

Combining multiple methods or data sources to strengthen the validity of findings

Answer

Triangulation increases confidence in research findings by corroborating results across different methods, data sources, or investigators.

Question 3

An analyst reviewing program data notices that participants who chose to enroll in a voluntary job training program had better employment outcomes than non-participants. What methodological concern is most relevant?

Accepted Answer

Selection bias—motivated individuals self-selected into the program

Answer

Selection bias occurs when individuals who choose to participate differ systematically from non-participants, making it difficult to attribute outcomes to the program itself.

Question 4

Which approach is used to assign dollar values to non-market policy outcomes (e.g., lives saved, environmental quality) in cost-benefit analysis?

Accepted Answer

Contingent valuation or willingness-to-pay methods

Answer

Contingent valuation surveys and willingness-to-pay methods elicit how much individuals would pay for non-market goods, enabling their inclusion in cost-benefit analysis.

Question 5

A researcher finds that cities with more libraries have lower crime rates. Before concluding that libraries reduce crime, the analyst should consider:

Accepted Answer

Whether a third variable (e.g., median income) explains both library prevalence and crime rates

Answer

A spurious correlation may exist if a confounding variable causes both the independent and dependent variables, creating an apparent but misleading relationship.

(PAC) Policy Analyst Certification Practice Test

(PAC) Policy Analyst Certification Practice Test

PAC Research Methodologies & Data Analysis 5