Term
|
Definition
| % of civilian labor force that is unemployed |
|
|
Term
| Bureau of Labor Statistics |
|
Definition
| agency in charge of employment numbers |
|
|
Term
| BLS interviews for monthly report |
|
Definition
| 1700 staff members interview 110,000 people |
|
|
Term
| PSU’s in 1990 redesign of Current Population Survey |
|
Definition
|
|
Term
|
Definition
| Primary Sampling Units = county or city |
|
|
Term
|
Definition
| Ultimate Sampling Units = 4 housing units |
|
|
Term
| At the second state USU’s are selected at random |
|
Definition
| every person 16 yoa and over are added to sample for the Current Population Survey |
|
|
Term
| Rotation of part of the sample = state one and state two |
|
Definition
| 16 different samples made = housing unit kept for 4 months = drops out for next 8 months = brought back for final 4 months |
|
|
Term
| Why the two stages and the rotation |
|
Definition
| saves money and helps estimate monthly change in unemployment and employment |
|
|
Term
| Threshold size of sub-sample in employment and earnings |
|
Definition
| 1000 estimates are done when sub-sample is below 50, no estimate is undertaken |
|
|
Term
|
Definition
| random sample formulas do not apply to cluster samples |
|
|
Term
| Which has more information Random or Cluster samples |
|
Definition
| random samples = person and neighbor have remote chance of getting into same sample |
|
|
Term
| Which has less information Random or Cluster samples |
|
Definition
| cluster samples = person and neighbor have more of a chance in getting into the same sample |
|
|
Term
| Half-sample method = split survey into 2 pieces |
|
Definition
| takes difference of estimates and this equals the SE = the average of the two is taken as the estimate of the characteristic to be estimated |
|
|
Term
|
Definition
| to know how sample was picked |
|
|
Term
| Bureaus’ SE for estimate in size of labor force is 5% smaller than for a sample random sample. WHY? |
|
Definition
| Because weighs are doing a good job |
|
|
Term
|
Definition
| Adjusting the weights helps correcting imbalances caused by chance variation and reducing sampling error. |
|
|
Term
| Is it possible to look at data and detect bias when it operates evenly across the sample? |
|
Definition
| NO. SE’s computed by the half-sample method will not pick up this kind of bias. |
|
|
Term
| About 3% of the monthly sample is re-interviewed by supervisors |
|
Definition
| and the supervisors reports are checked for incomplete of inconsistent entries |
|
|
Term
| Bureau’s SE for estimating the size of the unemployement is 50% worse than for a simple random sample why? |
|
Definition
| because the clustering hurts |
|
|
