![residuals in fathom 2 residuals in fathom 2](https://images.squarespace-cdn.com/content/v1/5efc8bebd631253a5a0f7263/1601236296886-TP36SH9MI4V5NWHGOR35/figure+2.jpg)
Correlations can be found simply by dragging (quantitative) attributes to both dimensions of a Summary Table. The Summary Table now has an “Add Five-Number Summary” option when adding new statistics. Three other useful changes to know about: 1. Works for collections and other objects such as graphs, summary tables, etc.ĭata > New Cases Data > Rename Collection Analyze>Sample Cases Analyze>Scramble Attribute Analyze> Collect Measures Analyze>Stack Attributes Analyze>Estimate Parameter Anlayze> Test Hypothesis Analyze>Estimate Parameter> Linear RegressionĬollection > New Cases Collection > Rename Collection Collection>Sample Cases Collection>Scramble Attribute Collection> Collect Measures Collection>Stack Attributes Object>New>Interval Estimate Object>New>Hypothesis Test Object>New>Linear Model> Simple RegressionĬan also “Collect Results as Measures” for Estimate, Test and Model boxes Easier to drag down the corresponding icon Or use the new Model icon May also “inspect” other objects like graphs, sliders or models Table > New Attribute Table > Rename Attribute Table > Sort Ascending Table > Sort DescendingĪssumes a Case Table (and usually an attribute) has been selected. Fathom 1.1Ĭomment Start a new Collection, Case Table, Graph, Summary Table, Interval Estimate, Hypothesis Test or Linear Modelĭata > New Attribute Data > Rename Attribute Data > Sort Ascending Data > Sort Descending The vast majority of the line-by-line changes given later arise from the menu differences that are summarized in the table below. New to Fathom2 are menus called Object and Collection. Gone from Fathom1.1 are the Display, Insert, Data and Analyze top menus. Most of the Fathom1.1 to Fathom2 changes relevant to this task have to do with the naming and organization of the menus. Rather we focus on the changes that affect those using Fathom2 with the edition of Workshop Statistics that was written for Fathom 1.1. What's the relationship between age and education?Ħ.Adjustments to the Fathom Edition of Workshop Statistics to accommodate Fathom2 Quick Summary: This is not meant to be a complete guide to differences and improvements in the latest version of Fathom.
![residuals in fathom 2 residuals in fathom 2](https://ars.els-cdn.com/content/image/1-s2.0-S009830042030635X-gr2.jpg)
Put differently, explain how age could be a confounder.ĥ. (In fact, there is a long ways to go.) Explain how some of this might be accounted for by age. There's still significant variation in income "unexplained". Is the fit "good"? What do the residuals say?ģ. Using graphics, describe the relationship between education and income.Ģ. Note that twins are labeled, somewhat arbitrarily, as L and H.ġ. Start up ARC and "load" the dataset twins.lsp. We'll do this together as a class, and then you can repeat on your own.Īctivity 2. from fathom, choose "export" to save the data as an tab-delimited file. What is the confidence interval? Is it probable that the "true" slope is 0? (And why use the value of 0 for our null hypothesis?) The difference is that now we include assumptions that allow us to generalize our conclusions to a larger population.Ī) What would this larger population be in this case? Do the data allow this?ī) Fathom does not give the typical t-test output, but instead gives a confidence interval for the slope. The "fit" is the same as the least squares. Under the "analyze" menu, you can select "Estimate Parameters" to fit an inferential regression line. What does this say? Are there any influential values?Ĥ. Does this look good to you? You can also make a residual plot (from the Graph menu, again). The systematic part of our model assumes that the relationship between systolic and diastolic pressure is linear. You will also get an equation and, for no extra money, an R-square term.ī) What do you make of the R-squared figure?ģ. Use this to super-impose a least-squares line. If you click once on the graph, the menu will change to provide a "graph" option. Does an assumption of a linear relationship seem reasonable?Ģ. You can choose to do this on either the "before" or the "after" values.ġ. Our goal is to determine the relationship between systolic and diastolic blood pressure.