I am working on the Assessment of Quality of Internet Financial Reporting (IFR) by firms in Saudi Arabia. For this purpose I will select a sample of firms that practice IFR. Their disclosure will be checked against the standard disclosure norms. If I prepare a checklist of the quality parameters for financial disclosure, what should be the ideal methodology to see the compliance of these firms and to test the main hypothesis that "Firms in Saudi Arabia practicing IFR comply with standard disclosure norm". I will probe further if there special norms for IFR. In that case I will use those quality parameters applicable for IFR.
ISSUE: Discrete data analysis
ACRONYM RE-DEFINED: To avoid confusion between IFR (internet Financial Reporting) and IFRS (International Financial Reporting), let's change IFR to QIFR (Qiality Internet Financial Reporting) in Saudi Arabia.
QUALITY PARAMETERS: Assume that you have a list of what is considered to be the components of quality reporting: Xi: (x1, x2, ..., xn). From this list, you have to set a threshold value for "passing quality reporting", i.e. 90%, 80%, 75%, etc. let's call that threshold theta (θ). Thus, QIFR > θ and anything less than θ is considered failing.
Under this standard we now have a binomial data set. If XQIFR > θ then score 1, but if XQIFR =
I find Microsoft Excel most efficient to perform this type of task. We can implement any test we want and put in any equation or formula as we wish to calculate. Canned software is limited to its original pre-packaged routines. Let me know if you need further assistance. Cheers.
Hi Namjul,
I am working on a paper that I try to answer this task. However, I have not finished it yet. I think you will find useful this paper: Comparison of two methods for measuring compliance with IFRS mandatory disclosure requirements, Tsalavoutas et al. 2010.
With regards,
Michail
Najmul -
You noted that you "...will select a sample of firms," but your sampling methodology is unclear. Whatever you do, I have one suggestion: stratify carefully so that characteristics may be isolated. If you want to look at each of the strata, that isn't exactly stratification, which is to help increase accuracy of overall inference, but the grouping process is still very important, perhaps more so. You can perhaps have an indication as to how differences in categorizations impact the data. But beware of spurious results that can occur. There should be theory to backup any differences you find. Theory first and then data confirmation.
If you categorize the data, you may even be able to treat them as response propensity groups. It depends upon your sampling and any estimation methodology. Also, if you do any regression studies, if that is possible, or descriptive statistics, categorizing your data will also be helpful. It is best if you can use specific characteristics to make these groupings/'strata,' but you might have to 'stratify' by size of firm.
I am not clear on this type of study, but I do think that sampling from population subgroups/'strata ' is probably helpful, regardless.
Cheers - Jim
Najmul - Perhaps Paul has a possible procedure for a confidence interval for one binomial set of data, with n=50, and you could also do the same thing with each "quality parameter" separately. - Jim
You can start using official authorities reports regarding the issue if available. I am not sure how it works in Saudi Arabia, but in most countries, as profits are subject of taxation, financial statements are checked by fiscal authorities, and they issue annual reports on that.
- Badges
- Science topic
- Similar topics
- Philosophy
- Philosophy Of Science
- Reasoning