Variables that influence validity and reliability

Within a previous posting, 'Reflexivity in Emancipatory Action Research', it was acknowledge that variables can emanate from 'perception of understanding', 'historical interpretation' and 'empirical knowledge' of individuals. (Horsley, 2009). 

Another variable to bear in mind is that of the type of learner involved in action research. Whether that participation be active or passive, through observation of, or reflection on. Be it statistical, quantitative or qualitative evidence based. Learner groups can be very different, even if some elements might show characteristics reminiscent of previous students. The overall diversity of participants can manifest very different results in research. 

Resources and strategies used for a specific programme of induction by one faculty, may not necessarily be considered generic for use for all faculties across college. Its success being affected by the variable needs and perceptions of different learners. However, the successful elements of a specific programme could become the building blocks that other departments expand upon and take ownership of. Showalter (2003) states:

'One of the best aspects of the work of teaching is that, unlike scholarship, it does not have to be original to be good. We can borrow ideas and methods from our colleagues and our predecessors, dead or alive: we can imitate, copy, and plunder in the confidence that our students will benefit from every good teaching technique we can put into action in our own classroom.' (p.9)

Furthermore Stringer (1999) states:

'Collaborative processes not only generate a sense of purpose and energy but also provide the means for the accomplishment of goals and the solution of problems and produce conditions that enhance participants' personal, social, and professional lives.' (p.189)

Therefore, to ensure validity for the purpose of this research, it is proposed that a control group be used to reduce variables. this will consist of three similar tutorial groups from across all three colleges:

Group A to follow normal induction programme

Group B to follow a specific two-week induction programme

Group C to follow the specific programme but with additional training given

Hopkins (2008) states:

'If the various threats to validity are not taken into account, then one cannot claim that one's interpretation is correct. The existence of possible sources of invalidity potentially offer plausible, rival interpretations to our findings when we do not account for them.' (p.140).

Another aspect for consideration, is that of reliability. To truly understand reliability in the context of educational research, a definition needed to be created. The result was as follows:

Reliable action research employs a process of methods that can replicate evidence which is constant in its equivalence over a span of time, if carried out on a comparable learner group.

McCormick and James (1989) state:

'Basically reliability is concerned with consistency in the production of results and refers to the requirement that, at lease in principle, another researcher, or the same researcher on the same occasion, should be able to replicate the same piece of research and achieve comparable evidence and results.' (p.188)

A second cycle of qualitative and quantitative questionnaires and statistical evidence would be desirable after dissemination of a specific induction pack across college in its entirety. If only for evaluation of its effectiveness, but more importantly to show credibility and stability of findings over a span of time.

References:

Hopkins, D. (2008) A Teacher's Guide to Classroom Research. Bershire, Open University Press. p. 140

McCormick, R. and James, M. (1989) Curriculum Action Research. (2nd Ed). London, RoutledgeFalmer

Showalter, E. (2003) Teaching Literature. Oxford, Blackwell Publishers. p. 9