Evaluation of Software - Stata

Evaluation of Software - Stata

Introduction

The purpose of this writing is to provide an evaluation of the statistical software which is used to do multinomial calculations and modeling. The version for which the evaluation is conducted is 3.1. The result from the use of this statistical software proposes the unique and distinguishing features which are incorporated within it for the easiness of students and also professionals like statisticians, econometricians, economists, policy makers etc. here in this paper only features are included for the purpose of evaluation. To provide a true and few view to the user for the performance of Stata; data is used and then a detailed comparison with other soft wares is given in order to understand the features and performance of the soft ware in comparison. In last the recommendation is given on the performance of Stata while keeping a comparative analysis of Stata with other statistical packages which are used as Social Science Computing Cooperative.

Discussion

The discussion is conducted firstly from taking a large data to check and present the performance of the Statistical software so that the whole evaluation can become creditworthiness.

Stata

Stata is a software package which is used for general purposes in many fields. Stata has multiple full range capabilities. It can do management of data, statistical and performance analysis, creation of simulations, graphic designing and customized programming. All these multiple features of Stata have enabled it to use heavily for research purposes. It has been penetrated in the academic circle of students and teachers and has been appreciated also through time saving and customized features of it.

Data

The data set was comprised through surveying method of young people who were rejected from providing their services to military due to failure in passing the exam termed as Armed Forces Qualification Test. The data was supposed to test on Stata and then for comparison on other statistical software's. For this purpose a contingency table was constructed. The table can be seen in Appendix A.

Comparison

In SAS; CATMOD procedure is used which is unlike Stata uses last level of response variable. Therefore, extra commands have to be added during computing. The parameters used by CATMOD are also different as they are used by Stata. To make the parameters identical with Stata, an extra coding is required to be done. If apart from technicalities; we do feature comparison then we come to drive the following result:

400 parameters can be fit in Stata, 150 in LIMDEP and none in SAS.

Specified value of dependent variable can be treated as a base category in SAS and Stata but this feature is absent in LIMDEP.

They all have an option for computation of conditional multinomial model.

Stata and SAS test hypothesis about all the coefficients which are present in the model but this feature is absent in LIMDEP.

Stata can test relative risk ratios while other two soft wares are unable to do it.

T statistics, z statistics and confidence intervals for parameters and odd ratios are all the features exclusively available ...