Q: What is the best way of designing a study that is assessing 3 or more companies?

Thank you for your question. A simple random sample, as suggested by you, is relatively uncomplicated and will reduce the potential bias associated with other methods.

However, it’s important to bear in mind that a reduced bias doesn’t equate to a bias-free study. For instance, since we’re surveying a particular number of companies here, even if all employees within these organizations are willing to participate, their views on personal development may be skewed towards the general view in the company, and if most employees’ values align with those of the company, we may not achieve the diversity we’re hoping for by surveying a larger group. Of course, bias is unavoidable in any experiment, but like every other limitation, this must be borne in mind, so that any claim made by the study is viewed within context.

While we can indeed assume that personal development affects each employee in the company and therefore the participation is likely to be high, we cannot guarantee this outcome, that is, we must also bear in mind the possibility that simple random sampling may be infeasible because of a lack of volunteers; a contingency plan to circumvent this (how can we sample if access to all employees is limited?) may be useful.

Finally, assuming that participation won’t be a problem, managing the time and costs of surveying each and every employee in all companies studied may be a challenging task. Such costs should always be weighed against the benefits, which in turn are a function of the scale and impact of the study itself as well as the advantages of the sampling method in terms of reducing bias and including a wider group of participants.

For the areas, convenience sampling, as mentioned by you, may be a relatively inexpensive and easy method, where we are likely working with companies who are readily available (thus facilitating our work). However, this method may hamper generalizability; for instance, if your sample consists of companies from a particular region that is not cosmopolitan, the results may not apply to firms located in more multicultural regions. Justifying your choice of region to journal editors/reviewers and convincing them that you haven’t over- or under-represented companies/employees in any area may require some work.

Also, since the method entails conducting surveys in a region where most companies are willing to participate, the relative lack of willingness of companies in other areas to participate may itself indicate a difference in their attitudes towards personal development and may thus be worth investigating.

Perhaps the strengths and weaknesses outlined above of both methods can be compared with those of alternative methods such as purposive and quota sampling, to help you arrive at a decision. You may also avail our Statistical Review service, under which a subject area specialist can help you ensure a robust study design, by helping you understand a pre-defined method to standardize your research, use appropriate tools and scientific instruments to ensure neutral results, determine the starting point of an investigation and a basis for reasoning, ensure a varied sample to arrive at an unbiased conclusion, identify and establish the problem to be solved, and outline variables that will play a key role in your research.

I hope you’re able to choose an appropriate method for your interesting research topic!