Statistics in Research Articles

Now that you are more familiar with data and statistical analysis including the selection of tests and familiarity with what the results mean, let’s look at a research article and see how we can apply what we have learned so far.
 

Common Research Article Structure

Abstract:

A paragraph long summary of the research including the context, research questions, methods, and findings.

Introduction:

Provides background on the research topic including existing research.

Methods:

Outlines the study participants, recruitment methods, data collection and statistical analysis.

Results:

Provides the findings of the statistical analysis, presents descriptive then statistical data and findings of significance.

Discussion:

The key findings, the meaning of the findings in the larger research picture and areas for further research. Limitations of the research are often included here.

 

The bulk of the statistical analysis and findings are in the results section of the article. Click here for an example of a research article. This article covers research conducted in South Korea on the relationship between the overestimation of body weight and lifestyle, dieting, and depression. There are several tables outlining the results of the statistical analysis and the p-values.

Using a simplified example of the relationship between scores on a scale measuring perception of risk, between pre and post-test among youth in the fictional Arizona Youth Program, here is what a table of results could look like:

 

Pre-test
(N=100)

Post-Test
(N=100)

P

Score
(1 = no risk, 2 = a little risk, 3 = some risk 4 = a lot of risk)

 

Item 1

2.7

3.4

0.03*

Item 2

2.8

3.2

0.065

Item 3

 3.0

3.7

0.04*

Item 4

2.4

3.0

0.07

Item 5

2.2

2.9

0.02*

*Indicates significance at the P <0.05 level

In the body of the article or report the data would indicate the type of test used (a paired t-test, for example), the p-value, and the effect size (if calculated). The p-value in the above table shows you that the relationship between the pre- and post-test. There was statistically significant improvement in three of the five items: items 1, 3, and 5. Information on the statistical analysis used, measurement tool (survey), and effect size would also be contained in a report.