Workplace Change and the Method Choice
A union at a food-processing plant reports that since a new automated packing line opened, 60 of 200 workers have been moved to night shift, injury reports have doubled, and agency staff now make up a third of the workforce. Management says productivity is up 14% and that 'workers were consulted'. The union wants research evidence before enterprise bargaining. A researcher has eight weeks and a small budget. She proposes a survey of all 200 workers, but the union steward warns that agency staff fear losing shifts if they criticise management, and that night-shift workers are hard to reach. Management offers access to its own incident records.
1. Compare two research methods the researcher could use and recommend one, justifying with reference to validity and ethics.
Read the worked answer
A survey of all 200 workers offers breadth and can quantify patterns such as injury experience and shift changes, giving statistical validity and reasonable reliability if the instrument is standardised. But it has ethical and validity risks here: agency staff may fear consequences, so responses may be biased toward management's account, and night-shift workers may be under-sampled. Semi-structured interviews with a purposive sample of day, night and agency workers would offer depth and allow trust to build, improving validity on sensitive topics such as fear of retaliation, but with 30–40 interviews it would be hard to generalise and time is short. A mixed design is strongest: a short anonymous survey plus 12–15 interviews, with confidentiality safeguards and recruitment through the union rather than supervisors. That addresses the power imbalance, which is the central threat to validity in this setting.
2. Explain how the researcher should handle management's incident records.
Read the worked answer
The records are useful because they are administrative data collected independently of the researcher, so they can corroborate or challenge worker accounts of doubled injuries. They improve validity through triangulation and are cheap in time. However, the researcher must ask how incidents were recorded, whether near-misses were included, and whether the definition changed when the new line opened - a change in recording practice could produce an apparent doubling. She should also seek access conditions in writing, ensure workers are not identifiable, and avoid letting management control which records are released. Using the records alongside worker accounts and the union's own figures gives a more defensible picture than any single source, and protects against the criticism that the research is one-sided.
3. The union steward says agency staff 'fear losing shifts'. Explain how this affects the research and what the researcher should do.
Read the worked answer
The fear creates a power imbalance between researcher and participant and between worker and employer, which threatens both ethics and validity. Ethically, participation must be voluntary and informed, with no way for supervisors to know who took part. Validity is threatened because agency staff may give socially desirable answers or decline, so the sample would over-represent secure workers and understate problems - precisely the group most exposed to the new staffing model. The researcher should recruit through the union and community channels rather than through supervisors, guarantee anonymity including in small cells, offer interviews off-site or by phone, and report agency staff findings in aggregate. She should also state these limits in her report so readers can judge the findings rather than overgeneralise them.