1

This is from an article I am reading recently. The authors compared information seeking results derived from different information needs. Here is how they reported comparison.While I know "p" means significance level, I cannot figure out what statistics methods they used, and what do "M" and "F" mean in the report. What metrics are they?

Information needs routed to search engines were rated significantly higher for contributing to knowledge (M = 4.85) than those routed to social networks (M = 3.73), F(1, 950.1) = 9.49, p < .01.

The article is: Oeldorf-Hirsch, Anne, et al. "To search or to ask: The routing of information needs between traditional search engines and social networks." Proceedings of the 17th ACM conference on Computer supported cooperative work & social computing. ACM, 2014.

Kathryn
  • 13
  • 2
  • It seems that $M$ is the mean of the sample, and $F(\cdot,\cdot)$ is the F-test's statistics with the corresponding p-value. See: http://en.wikipedia.org/wiki/F-test – stochazesthai May 25 '15 at 08:02
  • The $F$ refers to an F-statistic as might occur in an ANOVA or regression. Earlier in the text it says "We used a mixed-effects regression model", so that's presumably the model for which the results are being reported. Given that the first time $M$ occurs is in this context: "The sample ranged in age from 18 to 64 (M = 24, SD = 10.10)", it's likely to be a sample mean; that would fit with its use in the part you quoted if the means were being compared. – Glen_b May 25 '15 at 12:46

1 Answers1

2

I'm not an expert in statistical analysis, but it seems that it is ANOVA (analysis of Variance) results reported in APA style. M is the mean value of all the results, but instead of reporting the standard deviation for all the answers, the authors decides to analyze whether the results are statistically significant or not.

This type of analysis comes handy when you are dealing with human subjects, with lots of inter-subject variability and you wish to observe the effect of a treatment on them. Specifically, you are interested in seeing whether your treatment is effective on them or not (= whether you changed the mean score for statistical significant portion of participants or not).

In the case of this article, they provide the F-value to support that their treatment was effective on the subjects. Take a look at this article.

Kourosh Meshgi
  • 776