Improving Fruit Sorting Accuracy
First of all, make sure managers agree on which fruit meet their quality criteria.
International fruit markets often demand high quality packing criteria that need to be carefully communicated to packing shed employees. Crew workers can help improve the quality of packed fruit and increase the number of packs per bin if they better understand quality parameters.
This research report summarizes work carried out with three apple packing companies in Chile during the 2007 packing season. Our objective was to measure and improve reliability (consistency) as well as accuracy of the decisions of apple packing shed personnel—from management to packing staff. Two costly errors during sorting include discarding good quality fruit and packing bad quality fruit.
While the study could have been carried out in a number of agricultural enterprises, we chose apple packing sheds, where employees have to make many decisions quickly. After harvest, apples are brought to the packing sheds in bins and are sorted and packed according to quality standards. Packing decisions can be quite complex.
This study attempted to identify those individuals who were able to make accurate sorting decisions given a specific packing norm. Women are prevalently employed in this task, and were paid by the hour. The study involved several steps, including defining the criteria, verifying the criteria, training personnel, and testing personnel.
Defining the criteria. The packing shed client or managers determined what types of defects would be permitted in packed fruit.
Verifying the criteria. Managers and quality control personnel participated in an exercise in which they had to evaluate small samples of apples (25 to 50 fruit per sample), in terms of whether or not each apple should be packed, taking into account the predetermined criteria. Each member of the team was asked to evaluate the apples independently. After that, a conversation was facilitated during which each person could either defend their decisions or come to agreement with the others on how the apples should be graded based on the established criteria.
If management disagrees on which fruit meets high quality criteria, how can packing shed employees be expected to fare any better?
The goal was to identify at least two management or quality control individuals who had a good eye (were consistently 92% accurate) and would be able to train the line workers to accurately sort the fruit --according to the criteria.
The average accuracy of the management and quality control teams increased from 85 to 95 percent through this process. It should be pointed out, however, that individuals who did not improve substantially were eliminated from the teams.
Worker training. Packing line workers received training in two steps:
- a detailed explanation of the types of fruit damage and their causes; and
- hands-on exercises in which they would study, evaluate, and receive --feedback on decisions they made.
Formal testing. In each of the three packing sheds, samples (100 to 150 apples each) were numbered and spread over several tables. Participants were given a sheet of paper and pencil in order to note their opinion as to whether each apple in the sample should be packed. When finished, they turned in their completed sheet to one of the researchers, and were given a new blank sheet on which to --evaluate the next sample.
There were two samples per packing shed, and each subject was expected to evaluate each sample twice. Subjects were given about 20 seconds per fruit, but these times were reduced as they became comfortable with the testing process.
For each sorter, we wanted to get a reliability rating (how consistent they were, e.g., if they chose to discard apple number 14 the first time through, did they choose to discard or pack it the second time through?) and an accuracy rating (how their answers compared to the answer key).
Individuals who are not too sure of themselves are more likely to throw away good fruit. Furthermore, those who are not too sure of themselves may be seen discarding fruit in a packing shed line in order to seem busy.
We found great variability in terms of people's abilities to make correct decisions. Subjects ranged from 95 to 68 percent accuracy. We also found in this study that quality control personnel scored worse than packing shed employees in some instances.
Some of the benefits from this work include:
- Improving communication among management team members;
- Once standards have been developed, more accurately conveying those to packing shed employees; and
- Using this tool when deciding whether to hire workers or determining what type of job to give them. While we expect some employees to make major gains in terms of quality decisions they make, others will not be able to improve enough given the rate of speed required of them in making these subjective quality evaluations. Accurately identifying borderline fruit is likely to make a significant improvement to the bottom line.
Learn more about labor management at http://www.cnr.berkeley.edu/ucce50/ag-labor/