The Methodological Consultancy Process involves two parties: the consultant and the consultee/client. It is an interactive and creative activity in which a variety of methodological and statistical problems can come up. Frequently, clients start by asking simple statistical questions but at the end these questions may be only the tip of an iceberg. More and more elaborate problems of a methodological nature may be brought up.
Contents
Possible consultation setting(s)
(a) Help desk; (b) ‘Peer’ advice; (c) By telephone. These are by far the most difficult forms of consulting, because you are supposed to immediately provide a solution; (d) By e-mail. This form of consulting is efficient and appropriate for questions that do not require a long (theoretical or technical) explanation; (e) By Skype. Very nice new possibility, in particular if your client is far away or needs immediate help. Drawbacks: (i) You need quiet surroundings for yourself and your client, and (ii) if you are online a lot, you may be called unexpectedly; (f) Face-to-face without data analysis; (g) Face-to-face with data analysis; (h) Group meeting.
Phases of the consultancy process
Consultation session should have a beginning, a middle and an end. The session should begin with a clear re(formulation)of the problem and its context, then eventually proceed to its solution. At the end, a summary should be given and allocation of the responsibilities of the consultant and the consultees should be agree upon. Regarding the consultancy process, three phases in a sequence of consultation meetings can be found in the literature: (i) Problem statement and clarification phase, (ii) Problem solving phase, and (iii) Planning of the implementation of the solution.
A consultation process to be effective should ensure timely exchange of all relevant information between the consultant and the client. Consultations should start as early as possible in the research cycle so that it is meaningful to the implementation of the study. Moreover, the consultation process should be flexible, designed to fit specific situations and needs, carried out within a reasonable amount of time.
In the planning phase of a study, the consultant can ask questions on the design and the measurement instruments. He or she may suggest modifications to the design or recommend to do a pilot study to assess the quality of the instruments. On the other hand, if the study is already being carried out, the consultant can provide Advice (opinion) on the data collection and on the data quality. Problems such as handling of missing data or correction of confounding factors may come up in such cases.
Generally, consultancy questions are divided into four main categories,:
Design Questions
Many questions concerning the design of a study is for instance about Survey Design (Statistical survey), Web-Based Survey Administration, Increasing Response Rates, Weighting and Oversampling, Avoiding Social Desirability and Leading Questions, Research Design Issues, Sample Size Issues (Power Analysis or Statistical power), Focus Groups, Interviewing Techniques, Mixed-Method (qualitative/quantitative) Studies, Computer Simulation and Modeling,Computational Social Science or Multiple Criteria Optimization (Fuzzy Logic/Derringer method). Some of these questions may be hard to answer, especially when the problem is not clearly stated or when the consultant lacks the specific knowledge that is needed. In such cases, clients should be convinced that the question should be reformulated and in some cases to start all over again.
Questions on Measurement Instruments
Three kinds of questions on measurement instruments: 1) the choice of an existing instrument, 2) the construction and analysis of a new instrument and 3) the application of an instrument in a practical research setting.
Data Analysis Questions
The consultant may be confronted with a variety of questions on data analysis, for instance, what is the best way to analyze and handle missing data and/or outliers? or how do I determine how many factors I have to retain in maximum likelihood factor analysis?. Moreover, consultees have the opportunity to ask Advice (opinion) on how to conduct a variety of data analyses such as Univariate Statistical Analysis, Linear Regression, Multiple Regression, Logistic Regression, Analysis of Variance (ANOVA), Repeated Measures ANOVA, Analysis of Covariance (ANCOVA aka ANOCOVA), Variance Components (VARCOMP, Components of Variation, Gauge R&R), Mixed Models, or analysis of any of the experimental designs,. Many of these questions may look harmless but unanticipated problems can always come up.
Questions concerning Reporting
Often clients ask assistance on formulating methodological passages in order to publish their research. The consultant's response to this type of questions depends on the client's knowledge of the methods that were used.