Research Article Open Access

Biases from Poor Data Analyses

Tshepo Matsose1 and Solly Matshonisa Seeletse1
  • 1 Sefako Makgatho Health Sciences University, South Africa
American Journal of Applied Sciences
Volume 13 No. 10, 2016, 1033-1039

DOI: https://doi.org/10.3844/ajassp.2016.1033.1039

Submitted On: 10 July 2016 Published On: 5 October 2016

How to Cite: Matsose, T. & Seeletse, S. M. (2016). Biases from Poor Data Analyses. American Journal of Applied Sciences, 13(10), 1033-1039. https://doi.org/10.3844/ajassp.2016.1033.1039

Abstract

Non-statisticians with little knowledge in basic descriptive statistics tend to think that statistics field is limited to the content to which they are exposed. Many of them believe that a statistical package can augment the little Statistics knowledge they have. They often have a tendency to perform their own data analyses and do not even bounce it against Statistics experts for quality check. Many studies were concluded from data analyses performed by analysts who lack insight into statistical methods. Hence, results in some of their researches have flaws and distorted truths. The paper explains the defects in data analyses and research results that can be caused by influences in the data. Flawed research results may be caused when the data were not scanned for variations and other inconsistencies present in the data. Properly trained statisticians who also understand theories and methods of dealing with outliers can perform these analyses more effectively. However, many researchers fail to seek their advices. This study shows the extent of falsifications that contaminated data can produce and the massive loss to the factualness contained in the data.

  • 733 Views
  • 1,035 Downloads
  • 1 Citations

Download

Keywords

  • Data Variations
  • Information Falsification
  • Statistical Falsehood