These initial, “request for help” contacts are variable in their frequency, like an inhomogeneous Poisson process. They are also variable in urgency, importance, complexity, and familiarity. The requestor may be someone you have never worked with before. Or it may be someone you have been actively collaborating with.

It is critical for a good nonclinical statistician to make no assumptions about the request or the requestor before a first meeting is held. It is certainly tempting to worry about yet another unplanned challenge that will add stress to an already heavy workload. But there is always the opportunity to start or continue building a relationship, valuably contribute to an important scientific or business endeavor, or expand your knowledge and skills.

No requests for help should ever be turned down. After a request comes in, it needs to be dealt with in a timely manner. Our policy experience is to respond within the next business day. This results in a maximum of 2 full business days, given the modern nature of global time zones and 24/7 connectivity. The initial response may be no more than an acknowledgment that the request has been received, reviewed, and will be dealt with soon.

The minimum obligation is to provide an initial consultation, which invariably means one hour or less. Within our large company, requests occasionally come to us from outside our official scope of nonclinical pharmaceutical research and development, due to network and reputation. We owe it to our company, and as statistics professionals, to properly perceive and understand the need of the requestor without pre-judgment. Then a healthy dialogue can take place to come up with options which may or may not involve continued collaboration with a nonclinical statistician from the staff.

A request as previously described will typically turn into a project. Our definition of “project” here is that at least one more action needs to be taken beyond the initial consultation to fulfill this request and declare it as completed (Allen 2002). This determination of a project comes at the end of the initial consultation. One can optionally declare the actual discussion of the initial consultation as a project, if no further tasks are defined and agreed to for the nonclinical statistician to execute. If tasks for further collaboration are agreed to, the nonclinical statistician should proceed with defining and managing a project or projects within their organization system, stemming from the request. Projects can range to days to weeks to months to years.

We recommend an initial request be followed-up with a sincere desire to meet with a person and understand the entire context of the statistical needs. As mentioned earlier, this necessary level of understanding is impossible from an email or other communication channels. Except for rare, straightforward cases, such as explaining the difference between a standard deviation and standard error, replying to an email with a solution that you believe answers the question must be avoided. (Another category of exception is document reviews.) The prospective collaborator may believe their request is a simple question or questions, but virtually always it is not. All this can be clarified with an initial consultation. If the initial consultation can feasibly take place face-to-face, it should, especially for new or early relationships. In our modern world, this is not always possible due to geographical or time constraints.

The professionalism of a nonclinical statistician will start to reveal itself immediately to a requestor in how they handle the request and proceed to the initial consultation. As the saying goes, “first impressions count”, whether fair or not. Use phrases on behalf of yourself or your organization such as:

None of these phrases commit you to any obligations beyond the initial consultation. Even if they do provide background materials, you may not have the time to do more than glance at them, which is fine since they will not provide all the information you need. Your genuine interest will also be clearer to the requestor if you take the initiative to set up the appointment time and location that makes it convenient for them.

When you meet, allow your requestor and (potential) collaborator to provide all the necessary information in order to understand what is needed. Strive for a good conversation and healthy dialogue at all times. There are boundless references on how to do this, especially for the “listen first” and “seek first to understand” (Covey 1989) stages. It is easier said than done, of course, but this constant attunement is critical to the quality of the relationship.

The researcher requestor may not have a clear idea of how statistics can help. A key charge for the nonclinical statistician is to translate the scientific questions into statistical frameworks such as (but not only!) hypothesis testing or interval estimation, whether the request involves data at hand or involves the planning of a study. Key technical pieces involve understanding of experimental units, factors, endpoints, design, missing data, excluded data, etc. Appreciation and understanding of the underlying science is needed as well, since later evaluation and the quality of the interpretation of the data depends on it. The rest of this book offers a comprehensive canvas of the diversity of research & development areas where a nonclinical statistician can help. Good general traits for the consultation aspects of statistician can be found for example in Boen and Zahn (1982) or within Hahn and Doganaksoy (2011).

The researcher may have already completed some analysis of the data and seeks your review or your assistance with misunderstanding or limitations of the software. This is increasingly common due to the wide availability of software for storing and analysis of data. In projects that involve big data, collaboration with other quantitatively trained colleagues is essential. For now, we mention that open receptiveness to ideas about statistical methods and data analysis methods is essential. If the initial consultation reveals that the project will be a more 1-to-1 or 1-to-small-group collaboration, the nonclinical statistician will have dominant control of the choice of statistical methods. If the nonclinical statistician is part of an interdisciplinary team with other quantitative colleagues, then confidence, but not arrogance, is needed to mutually choose appropriate data analytic techniques. Depending on experience and the complexity of the problem, one might find themselves unsure at all what statistical methods will be needed. It is perfectly all right to say “I don’t know” or that “I don’t have experience,” and also assure your requestor that you have the ability to search for accurate and sensitive approaches from literature, colleagues, etc..

Completion of the initial consultation requires a clear definition of what is needed and when it is needed, particularly if an urgent business milestone is approaching. Here is where good negotiation skills on part of the nonclinical statistician are a must. Besides the people present in the initial consultation meeting, there are other considerations of stakeholders that are not present that will be taken into account, and these must be disclosed. It may be that a presentation to management is coming up next week and one or a few slides of statistical content are needed to cover evaluations and interpretations of a key endpoint. A more comprehensive report can wait until later, or eventually may not be needed at all. Everyone is busy, and we have found universal recognition by researcher colleagues that the ideal situation of fast as possible delivery of results and perfect work is not attainable. One solution may turn out to be for the nonclinical statistician to serve as an advisor for the researcher to continue doing their own analysis with their own software package. As Allen (2002) describes, every project can always be done better with more time and/or information. Allen furthermore advises, and we paraphrase here: “Lack of time is not the major issue. The real problem is a lack of clarity and definition about the project and what is required.”

One more topic we wish to mention here is the “curse of knowledge,” and its effect on interactions, starting with the initial consultation. As Heath and Heath (2007) discuss, “once we know something, we find it hard to imagine what it was like not to know it.” As with all professional fields, constant self-checking of this natural human tendency is needed in order to prevent obstacles in communication by a nonclinical statistician to a non-statistician. Statistical jargon will be needed when a statistical expert verbally discusses, writes, or presents statistical concepts related to problems and solutions. Awareness to simplify as much as possible with the listener’s perspective in mind will be beneficial to effectively communicate complexity and to build credibility. Practice of this behavior is needed all the way through and to the end of the project in order to continue reception of its benefit.

The initial consultation also unequivocally identifies the lead nonclinical statistician contact for the project. Usually this is the statistician who assumes the lead for the initial consultation in the first place. Exceptions to this would include professionals in the early stage of their nonclinical statistics career, where a more senior colleague will assist. In more complicated, longer term projects, this lead statistician may also need to coordinate with other nonclinical statisticians to perform portions of the work. We feel the taking on of full professional responsibility and accountability for a project is a vital component of the nonclinical statistician who is identified as the lead. It is an end-to-end endeavor.