GO TO THE FULL PROTOCOL:
PDF or HTML at Wiley Online Library
Basics of Statistical Analysis
Publication Name:
Current Protocols Essential Laboratory Techniques
Unit Number:
Appendix 4B
DOI:
10.1002/9780470089941.eta04bs03
Online Posting Date:
June, 2010
Abstract
The goal of this appendix is to introduce basic methods in graphing and data analysis and explore some fundamental concepts in statistical reasoning. Three examples of data-analysis problems relevant to molecular biology are used to illustrate methods covered in a first statistics course, including the two-sample t test, simple linear regression, and chi square tests for goodness of fit and contingency table hypotheses. The appendix also explores the selection and interpretation of appropriate summary graphs for these analyses, including the use of error bar plots, scatterplots, and bar charts. In addition, a number of key terms and concepts are introduced and explained in the context of the three example problems, including summary statistics, sampling variation, the standard error, null hypothesis testing, the use of test statistics, and the interpretation of p values. Curr. Protoc. Essential Lab. Tech. 3:A.4B.1-A.4B.22. © 2010 by John Wiley & Sons, Inc.
Keywords: data analysis; graphing; standard error; two-sample t test; linear regression; chi square; goodness of fit; contingency table
Figures
-
Figure A.4B.1 Summary graphs for the three example problems. (A) Example 1 is a comparison of continuous responses between categorical groups, best summarized with an error bar graph (mean ± one standard error). (B) Example 2 is a test of the relationship between two continuous variables, best summarized with a scatterplot. (C) Example 3 is a comparison of categorical outcomes between categorical groups, best summarized with a clustered bar graph. -
Figure A.4B.2 Three hypothetical comparisons between two mutant lines, illustrated with graphs of the means ± one standard error. (A) High signal and low noise. (B) High signal and high noise. (C) Low signal and low noise. -
Figure A.4B.3 Simple linear regression analysis for Example 2. (A) The best fit regression line and equation. (B) The null and alternative hypothesis, with a visual illustration of signal and noise.
