Article Summary: Evaluating the Econometric Evaluations of Training Programs with Experimental Data by Robert J. Lalonde

Posted: August 27th, 2021

Student’s Name

Instructor’s Name

Course

Date

Article Summary: Evaluating the Econometric Evaluations of Training Programs with Experimental Data by Robert J. Lalonde

The article by Robert J. Lalonde: Evaluating the Econometric Evaluations of Training Programs with Experimental Data conducts a comparative analysis of the impacts on trainees’ earnings in an employment program. The study was implemented as an experiment through which participants were randomly selected and assigned to study treatment. The study adopts a control group to help assess the estimates obtained during the experiment. The study compares the results attained during a field experiment against those obtained by econometrician estimates without conducting an experimental study. It borrows some econometric tests’ ideas to assess the extent of sensitivity of the results against the specifications of an alternative econometric analysis based on advanced tools (Lechner, Michael, and Pfeiffer 117). The target training program is the National Supported Work Demonstration (NSW), which was implemented temporarily to train disadvantaged workers with the necessary skills to improve their work and hence performance. The target population included both women and men, school drop-outs, ex-drug addicts, and ex-criminals randomly distributed into treatment and control groups (LaLonde, 608). The model included random assigning of qualified applicants to train the selected employees. The treatment group was assigned with all the requirements while the control group was left to fend for themselves (LaLonde, 609). Thus, each participant in the treatment was given a job 9 to 18 months based on the sites they were assigned to work at a minimum wage with a promise for an increase depending on individual work performance. Besides, they all had access to a counselor where they would present their grievances as a team.

After every nine months, data on salaries and demographics were collected from both the control and treatment groups. Post-study inter-views accompanied this. However, it was noted that most participants failed to complete post-study interviews, which affected the overall results of the study (LaLonde, 619). Subsequently, about 72 percent completed the interviews in the treatment group, while only 68 percent of the control group participants completed the interviews. Further, assessing the average earnings from the random assignment for both groups shows it was the same both before and after training (LaLonde, 688).  Thus, there were no statistically significant differences between the two groups regarding individual employee characteristics, hours of training, and earnings. However, across gender, it was established that female participants’ average earnings before the NSW program were higher than after the program. At the same time, the study established that the econometric procedures did not impact the experimental estimates. The examination of non-experimental results reveals completely different results in each case.Thus, implying that a difference exists between experimental and non-experimental analysis when it comes to employment policy evaluation. Hence, the article also raises the issue with econometric specifications applied in assessing labour markets, tax applications and other economic policies affecting employment market since they still require to be evaluated to ascertain accuracy of their results.

In response to the study’s question, the experimental assessment establishes a tremendous difference between non-experimental and experimental results. The study attributes these differences to potential biases that are prone to non-experimental methods compared to experimental ones. For instance, non-experimental methods often produce positive results for female participants while negative for male participants when assessing the impact of training and employee earnings (LaLonde, 813). Therefore, it is advisable to check econometric evaluations using experimental studies before they are fully adopted into policies.

Works Cited

LaLonde, Robert J. “Evaluating the Econometric Evaluations of Training Programs with Experimental Data.” The American Economic Review, vol. 76, no. 4, 1986, pp. 604–620. JSTOR, www.jstor.org/stable/1806062. Accessed 18 Oct. 2020.

Lechner, Michael, and Friedhelm Pfeiffer. Econometric Evaluation of Labour Market Policies. Heidelberg: Physica-Verlag HD Imprint Physical, 2001. Print.