The question is about assessment of skewness and kurtosis to carryout a research in construction management through SPSS software. The purpose behind the question is to know the permissible values of skewness and kurtosis with proper reference.
Your query is a bit open ended. What are the parameters that are of interest when you refer to construction management? Is it time to completion? Cost? Probability of a successful product (vs. one that fails, like the Tacoma Narrows bridge)? Or, something else entirely?
Skewness is a measure of asymmetry in a distribution, and kurtosis is a measure of peakedness and heaviness of tails of a distribution; they are successive distributional moments after variance. To my knowledge, there are plenty of guidelines as to what "acceptable" levels of skew and kurtosis might be; these are usually given in the context of presuming that data follow some sort of specific distributional shape, such as Gaussian ("Normal").
However, in a procedure which assumes, let's say, a normal distribution, the specific threshold for skew and kurtosis beyond which type I or type II errors occur at "markedly" more than the nominal levels is situation specific (the specific variables, the sample and target population, the sample size, the procedure, etc.),
So, while there may (and probably are) guidelines that have been offered within the domain of construction management, my opinion is that the only way to be completely certain is to: (a) know the variables and context; and (b) run simulations to determine probable impact on the validity of subsequent decisions.
Thank you so much sir David Morse the answer is really helpful but still I did not get the permissible limits I have to mention these in my work. I am working on integration in construction projects. I have done these tests skewness and kurtosis in SPSS and I have got certain values but I don't know whether the values are acceptable or not.
Skewness is the degree to which a statistical distribution is deviant from symmetrical shape. Kurtosis is the degree of peakedness of the distribution. These are measured relative to the Gaussian distribution for which the values are 0 and 3. The Gaussian distribution will be a point of reference in the family of statistical distributions due to its symmetry and balance. In terms of construction management this distribution shall continue to be a reference point in many respects. Limits may not be easy to fix but reasonable values should do for many situations.