CALIFORNIA STATE UNIVERSITY LONG BEACH
PPA 696--RESEARCH METHODS
BINGHAM & FELBINGER CH. 15
  1. BIBLIOGRAPHIC INFORMATION
    1. Author: A. Rees
    1. Title: An Overview of the Labor-Supply Results
    1. Source: The Journal of Human Resources, 9, 1974:158-180.

      2.  
  1. SUMMARY OF THE RESEARCH
    1. PROBLEM STATEMENT:
What is the effect of unearned income on labor force participation?
    1. BACKGROUND:
Part of an experiment in New Jersey and Pennsylvania
    1. HYPOTHESIS:
Payment of unearned income will reduce labor supply by 10-15%
    1. MEASUREMENT OF VARIABLES
      1. Dependent variable: 1) Labor force participation rates; 2) employment rates; 3) total hours of work; 4) total earnings
      1. Independent variable(s): Payment of unearned income
      1. Control variable(s): health status; ethnicity; normal hourly earnings; pre-program values of dependent variables;
    1. RESEARCH DESIGN:
A quasi-experimental factorial design; eight different income payment plans; pre- and post-tests and a control group.
Group T1 Program T2
G-1 to 

G-8

 O1 X O2
Control 

Group

 O1 O2
 
    1. SAMPLING:
Non-random recruitment of low-income families in New Jersey and Pennsylvania, which were then randomly assigned to one of the 8 payment plans or the control group
    1. INSTRUMENTATION:
Data were gathered by income reports submitted by the families
    1. DATA COLLECTION/ETHICS:
Data were collected every four weeks from experimental families; less often from control families
    1. DATA ANALYSIS:
Percentage differences; regression used to estimate means for experimental versus control groups; two-tailed t-tests used to test for statistical significance of group differences.
    1. CONCLUSIONS:
A negative income tax will result in a reduction in a statistically significant but not very large labor force participation.
  1. CRITIQUE
    1. Possible Threats to Internal Validity
      1. History:
The results were probably affected by three events: 1) a change in the AFDC program in New Jersey to include unemployed fathers; 2) the payment of higher AFDC benefits than the experimental treatment; and 3) the subsequent cutoff of AFDC eligibility for unemployment fathers. There was substantial switching back and forth between AFDC and the experimental treatment plans. There were also economic swings, and may have been other events over the three-year period of the experiment.
      1. Maturation:
n/a
      1. Testing:
The knowledge that the experiment had a limited 3-year duration may have affected the behavior;
      1. Instrumentation:
The experimental group received more training on how to report gross earnings than the control group;
      1. Regression Artifact:
 
      1. Selection bias:
Volunteers were recruited for the study, who may have had different motivation than non-participants
      1. Experimental Mortality:
No data given; only families with complete data were used, which suggests that some families were dropped due to incomplete data
      1. Design contamination:
Families could obtain other benefits at the same time
    1. Possible Threats to External Validity
      1. unique program features:
      1. experimental arrangements:
The effects of the program cannot be disentangled from the other benefits that the families were eligible to collect at the same time, such as AFDC, Food Stamps, etc.
      1. other threats: