![]() Linearity is tested by analysis of variance for the linear regression of k outcome observations for each level of the predictor variable (Armitage, 1994): Generalised replicate analysis is best done as a part of exploratory modelling by a Statistician. Generalisations of this method to models with more than one predictor are available ( Draper and Smith, 1998). deviations from the regression line (residuals) have uniform variance.deviations from the regression line (residuals) follow a normal distribution.Y replicates are a random sample from a normal distribution.If your data are arranged so that each repeat of Y is in a separate column then use the Rotate Data Block function of the Data menu on your Y data before selecting them for this function. row 4 of x should relate to column 4 of Y repeats). Please note that the repeats of Y observations should be arranged in a worksheet so that each Y column contains repeats of a single Y observation that matches an x observation (i.e. The number of Y replicate variables which you are prompted to select is the number of rows in the x variable. Replicate Y observations should be entered in separate workbook columns (variables), one column for each observation in the predictor (x) variable. This enables you to run a test of linearity and thus justify or refute the use of linear regression in subsequent analysis ( Armitage and Berry, 1994). This function can be used to test that assumption of linearity.įor studies that employ linear regression, it is worth collecting repeat Y observations. The standard analysis of variance for simple (one predictor) linear regression tests for the possibility that the observed data fit to a straight line but it does not test whether or not a straight line model is appropriate in the first place. This function gives a test for linearity in a simple linear regression model when the response/outcome variable (Y) has been measured repeatedly. ![]() Menu location: Analysis_Regression and Correlation_Grouped Linear_Linearity. Open topic with navigation Linearity with Replicates of the Outcome (Y)
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |