Download presentation

Presentation is loading. Please wait.

Published byDestiny Long Modified over 3 years ago

1
DataToText: A Consumer-Oriented Approach to Data Analysis David A. Kenny University of Connecticut University of Connecticut http://davidakenny.net/doc/aps10.ppt http://davidakenny.net/doc/aps10.ppt http://davidakenny.net/doc/aps10.doc

2
Data Analysis for Methodologists Begin with a hypothesis that is embedded within a statistical model. Develop a research and measurement design to estimate the parameters of that model. Gather data; test model assumptions; and estimate the models parameters. Choose the best model.

3
Data Analysis for Practitioners What do I have to click in SPSS? What from my SPSS output do I put where in my results section?

4
Consumer versus Industry Perspective For data analysis, quantitative psychologists are the industry or at least part of the industry. The spirit of DataToText is an attempt to get us to become more consumer- oriented.

5
DataToText Project Have the researcher tell DataToText what is the research question. DataToText performs the requisite analyses. DataToText gives the results from those analyses: computer output a written description

6
Example 1 http://davidakenny.net/moderate.htm Moderation Analysis Livi et al.: The effect of Noise Sensitivity (X) on Stress (Y) is moderated by the Need for Cognitive Closure (M)

7
Moderation Example Syntax Gives results in words, tables, and pictureswords tablespictures Tests assumptions Gives warnings Specialized output

8
How Positively the Wife Sees the Husband Wife Satisfaction Example 2 Husband Satisfaction Actor Wife Actor Husband Partner Husband Partner Wife How Positively the Wife Sees the Husband http://davidakenny.net/ddt/apimd.htm SyntaxSyntax PicturePicture Specialized Output

9
Advantages and Disadvantages

10
DataToText is Mindless! Thought and intelligence is needed for: What analysis to do The execution of the analysis The interpretation of the analysis

11
High school student Jenna Smith randomly throwing variables into the macro.

12
Researcher Brad Anderson not knowing anything about testing moderation.

13
Definitely bad, probably terrible. However, it might be better than what would have otherwise been done.

14
Additionally Sometimes warnings may alert the user that the wrong analysis was done. For example, both macros provide a warning if a dichotomous variable is analyzed.

15
DataToText is Mindless! about: What analysis to do The execution of the analysis The interpretation of the analysis

16
Data Analysis Requires Thought Not all problems have a flow- chart structure CFA ARIMA modeling

17
However Some problems have a flow-chart structure (though we might disagree some about that flow- chart). For many analyses we do have explicit or implicit standards for reporting of results.

18
Also Keeps everything straight. Avoids errors. DataToText while it may typically fail to perform the best analysis, it might often create a better analysis than that done by even a skilled data analyst. Provides warnings. Makes assumptions explicit and can provide statistical tests of some of them.

19
Warnings DataToText issues warnings: A dichotomous outcome variable Outliers Colinearity Low Power

20
Assumption Testing Is Like Flossing Something we know that we should do but do not do as often we should. DataToText can test certain assumptions level of measurement distributional outliers

21
DataToText is Mindless! about: What analysis to do The execution of the analysis The interpretation of the analysis

22
The Researcher Needs to Understand the Results Good data analysis is more than doing the right analysis; the researcher must still understand the meaning of the results.

23
However Because DataToText provides a verbal summary of the results, some researchers might better understand their results by using the approach.

24
Moreover Output produced by DataToText may be more intelligible to the reader and so although the data analyst may not understand the results, the reader may be able to!

25
Packages vs. Open Source Packages PASW ( née SPSS), SAS Accessible but costly Updates sometimes without backwards compatibility Open Source: R

26
Feedback Please let me know what you think. david.kenny@uconn.edu http://davidakenny.net/doc/aps10.ppt http://davidakenny.net/doc/aps10.doc

27
Special Thanks to My Lab Thomas Ledermann Amanda Snook Randi Garcia

Similar presentations

OK

Skewness and Curves 10/1/2013. Readings Chapter 2 Measuring and Describing Variables (Pollock) (pp.37-44) Chapter 6. Foundations of Statistical Inference.

Skewness and Curves 10/1/2013. Readings Chapter 2 Measuring and Describing Variables (Pollock) (pp.37-44) Chapter 6. Foundations of Statistical Inference.

© 2017 SlidePlayer.com Inc.

All rights reserved.

Ads by Google

Ppt on conjunctions for grade 5 Ppt on electricity in india Ppt on p&g products brands Ppt on cross-docking benefits Ppt on chapter management of natural resources Ppt on ten figures of speech Ppt on indian culture and festivals Ppt on five monuments of india Free download ppt on railway reservation system Ppt on water scarcity articles