You started your work on a sound footing, however for complimentarity please check this url: http://ijbssnet.com/journals/Vol._2_No._9_[Special_Issue_-_May_2011]/8.pdf

LITERATURE REVIEW: First thing first, conduct a thorough literature review to see what other researchers had done on this subject matter. Is there something missing or gap in the literature? This literature review serves several purposes: (i) it allows us to have an over all picture of the subject matter, (ii) what variables (Xi) had been used as explanatory factors, (iii) what contribution can the new research (your Ph.D. thesis) contribute to the body of knowledge.

VARIABLES:

AKAIKE CRITERION: A research that "over identified" explanatory variable is neither robust nor attractive. The more explanatory variable one uses the less powerful is the proposed model. As a preliminary testing, run a small sample of pilot test and then implement Akaike criterion test to see whether the proposed model is over identified. In this case, the following had been identified as Xi: (1) Competition, (2) strategy, (3) size, (4) affiliation, and (5) listing status.

X1 may be quantitative and scale survey may be employed. This may be a single-factor analysis.  

X2 through X5 appear to be discrete, i.e. either ordinal or nominal data type---logistic and probit regression may be appropriate for these factors.

SCALE: This scale (1,2,3,4,5) appears to be not effective. It may be better to consider (0,1,2,3) where 0 = none, 1 = low, 2 = medium and 3 = high. As for the intensity, you may use correlation coefficient as the measure and then employ appropriate test statistic to verify their statistical significance. Try to avoid to embed the level of intensity inside the scale by making the scale extensive, i.e. (1,2,3,4,5).

ACCOUNTING PRACTICE: Check the local law where these companies are registered. You speak of "listing" in the stock market. In that case, accounting practice is a straight forward matter. All listing companies are required by securities regulation to adopt international standard in accounting practice, i.e. IFRS or IAS equivalence. For non-listing companies, their either adopt the same practice as listed companies or "else." If "else", it should considered below standard---if so, you may want to company their operational performance, governance, etc. with the listed group.

DEPENDENT VARIABLE: The dependent variable here is too simplified: "accounting practice." You might what to have sub-issues, i.e. Y1 = accounting practice, Y2 = operational performance, Y3 = governance, etc. In so doing, you have more things to discuss at the thesis defense.

I hope this has been helpful. Cheers.

actually this is a very interesting topic. I already made a similar research 2 years ago. the way i used was to make 0s and 1s to denote the use of the management accounting techniques like ABC, BSC, and EVA. the second method was through a questionnaire to review the acceptability and the applicability of the different strategic management techniques on the sample selected.

I very much agree with Paul Louangrath. A thorough search in related or prior empirical studies can guide you in the selection, measurement and definition of relevant variables.

Again, the concepts, models and theories underpinning management accounting practices can also guide the selection of variables.

Finally, remember that there is nothing like a "perfect research". Every research work may have some issues. However, it is your ability to provide adequate justification supported by relevant literature on the methodology that reduces the biases associated with your research.

Las primera variables son de característica cualitativas y la puedes medir con una scala likert aumentando la dimensión a 7 puntos.  "La competencia, la estrategia de organización, la cultura organización, tamaño, afiliación multinacional y del estado"  debes desarrollar preguntas y darle un valor ala respuesta a obtener y correlaciones a todas con la practica de contabilidad de gestión que sera una variable cuantitativa que demostrara el numero de instituciones  estudiadas; los resultados económicos, las sanciones o reconocimiento son otras variables a ser consideradas. Finalmente el aporte de impuesto ala renta sera el resultado de una buena o mala ''Prácticas de Contabilidad de Gestión 

.

Dear Oluyinka,

I see you have an interesting approach.

In any case, as you say that your “original intention is to use primary data” I also subject you to try (maybe later) to research in Accounting History.

You could see the article by Salvador Carmona, Mahmoud Ezzamel and myself entitled “Accounting History Research: Traditional and New Accounting History Perspectives”. Published in: De Computis, Spanish Journal of Accounting History. 2004; 1: pp. 24-53 (http://www.decomputis.org/).

If you have any question you could write me ([email protected]) without any problem.

Best wishes,

Fernando.