The Veritas Blog » Lessons and Seminars

Statistics: On Multiple Regression

admin — Tue, 31 May 2011 17:00:10 +0000

For the final installment in this series on introductory statistics for the social sciences, Kyle T. begins discussing multiple regression, a statistical test and topic that can be subject to much more advanced investigation by those students who are interested in learning more about this subject. In short, multiple regression extends regression in a way similar to how factorial AnoVa extends AnoVa; multiple regression allows prediction based on a number of different, independent variables.

We hope that you’ve enjoyed this overview of statistics, geared for social-scientific students and research, this Spring. We’re excited now, leaving Memorial Weekend and entering June, to be transitioning into a new series, geared for premedical students, preparing for the MCAT this Summer. Stay tuned, eager minds; the Veritas video series will return anon!

Warmly yours,

The Veritas Team

On Multiple Regression:

Statistics: On Regression

admin — Tue, 24 May 2011 17:00:21 +0000

As promised, this week Kyle T. addresses regression, a more complex correlation model that does allow for interpretations of causality. However, despite its relative complexity, regression is still just, as Kyle says, “glorified slope-intercept”! Following this logic, you’ll be up and running these analyses in no time.

Hoping you’re enjoying the Spring!

The Veritas Team

On Regression:

Statistics: On Correlation

admin — Tue, 17 May 2011 17:00:05 +0000

Robust like the non-parametric statistics that Kyle T. discussed last week, correlational analyses, Kyle’s topic this week, provide the raw materials for the much vaunted regressive analyses, his topic next week. In essence, as Kyle explains, correlations generally assess how two sets of patterns of data line up next to one another; and, as he emphasizes, however they line up does not say anything about however they may arrive in that relationship: Correlation does not describe causality.

However, the two are related; to learn how correlation and causality do connect, be sure to check back in on our next installment in this series: Regression.

Warmly yours,

The Veritas Team

On Correlation:

Statistics: On Non-Parametric Statistics

admin — Tue, 10 May 2011 17:00:08 +0000

This week Kyle T. addresses the other major type of statistics: non-parametric statistics. As he explains, non-parametrics allow for more robust testing than parametrics, since only non-parametrics can test with non-normal distributions and with ordinal (or less) dependent variables. However, as he adds, non-parametrics’ robustness is counterbalanced by their loss of statistical power; the “collapse” of levels of granularity or fineness within the information gathered about the groups being tested eliminates much of the incisiveness of the tests’ results.

Next up: Correlation!

Warmly yours,

The Veritas Team

On Non-Parametric Statistics:

Statistics: On Factorial AnoVa

admin — Tue, 03 May 2011 16:41:30 +0000

This week Kyle T. expands on the basic AnoVa that he’s presented over the past two weeks and shows how it may be used to test groups that differ one more than one categorical variable. This expansion is called the Factorial AnoVa and can be used in between-subjects, within-subjects, and mixed research designs.

Next week: Non-Parametric Statistics!

Statistically yours,
The Veritas Team

On Factorial AnoVa:

Statistics: On Analysis of Variance (AnoVa – Part Two)

admin — Tue, 26 Apr 2011 17:00:47 +0000

Wrapping up his explanation of the AnoVa, Kyle T. this week presents his third way of understanding the intention of the statistical test by graphically modelling distributions to be compared: By looking at distributions as representations, the comparison becomes clearer than the abstract concepts in consideration.

See you in May!

The Veritas Team

Statistics: On Analysis of Variance (AnoVa – Part One)

admin — Tue, 19 Apr 2011 19:57:14 +0000

This week, as promised, Kyle T. introduces the extension of the t-Test, the Analysis of Variance (AnoVa), meant for comparisons among more than two participant groups. His explanations of the AnoVa tie onto the heart of the test, discerning the variance within groups from the variance between groups; he turns to information theory for its descriptive analogy, educing a signal from noise.

Next week: A brief continuation of explaining the AnoVa, including graphs!

Warmly yours,

The Veritas Team

Statistics: On t-Testing (Part Two)

admin — Tue, 12 Apr 2011 16:00:24 +0000

Continuing last week’s discussion of the t-test, Kyle T. describes the Paired or Dependent Samples t-Test and the ways in which it relates to its already presented sibling, the Independent Samples t-Test. Though this relationship is not sequential as have been the relationships between the other previously presented tests, the two tests nevertheless logically support one another and continue the progression of complexity in statistical analyses that he’s been developing so far.

Next week: AnoVa!

Hoping that you’re enjoying warm weather this Spring,

The Veritas Team

On t-Testing (Part Two):

Statistics: On t-Testing (Part One)

admin — Tue, 05 Apr 2011 16:00:46 +0000

Hey, statistics students and aspiring social scientists; it’s t-Test week!

This week Kyle T. presents the basics of the t-test, the most foundational analysis in inferential statistics. Thus, this test and its consequential relatives will be the analyses that you conduct most often in your own research. Exciting!

Happy April,

The Veritas Team

On t-Testing (Part One):

Statistics: On Z-Tests and Sampling Distributions

admin — Tue, 29 Mar 2011 16:00:02 +0000

As promised, this week Kyle T. presents the first of several foundational mathematical tests, used in statistics to compare populations and samples. This week’s test is the z-test and is the most basic analysis conductible in descriptive statistics. For those standardized-test takers of you out there, this test parallels the analyses that the testing companies conduct in order to calculate your scaled scores from your raw scores.

Next week: t-Tests!

-The Veritas Team

On Z-Tests and Sampling Distributions: