The ball position factor should be analysed in a closer way in order to find the optimal position. To sum up, the shooting height factor has low influence compared to the other factors and can be taken out of the analysis. Using repeated measurements for each of the categories, once can compute mean confidence intervals. The smallest effect is for the factor shooting height with 277 for D1 and 257 for D3. In this example it is the factor ball position with a distance varying between 331 for B1 and 200 for B3. The factor with the highest variation has the biggest effect. For a given factor the mean distance for each of its categories is displayed and connected by a line. These charts show the effect of each factor on the launch distance of the projectile. The means for which the distance is the greatest are associated with the factors A3, B1, C3 and D1. If we then look at the means charts, we can quickly notice that we find the same results as the optimization. We can see that the R² = 0.894, which shows that the ANOVA describes the data very well. Then all the results of the analysis of variance are displayed, starting with the goodness of fit coefficients. This one shows us directly that the combination allowing to maximize the distance at which the projectile is launched by the catapult is the following: A3, B1, C3, D1. Interpreting the results of a screening design analysisĪ first table is displayed, providing a summary of the responses optimization. Once you have clicked on OK, the computations start. In the Options tab, check the response optimization option and select the associated table in the design sheet. Be careful to select all the design columns (see below). Select, in the general tab, the column with the results of the experiments as well as the experimental design table. The second is to click on the DOE button on the ribbon and select "Analysis of a screening design".This way, the a pre-configured dialog box will open ready to launch the computations. The first is simply to click on the button "launch the analysis" located below the design table.You have two options to analyse the design. Setting up the dialog box for the analysis of a screening desing In the demo file used for this tutorial, the results are already present, they have been highlighted in yellow to identify them quickly. Select the option to maximize the response in the optimization table, which will find the best combination of factors allowing the catapult to launch the projectile the farthest. Let's suppose that the experiments are carried out and the results are entered in the appropriate column of experimental desig table. Below, the response optimization table is displayed and will be automatically filled after entering the response results. Then the design is displayed, in which the last column corresponds to the response results and must be filled with the results of the experiments. Output of the experimental designĪ table with all the information related to the factors is displayed. Select the Latin Square design by clicking on the "Select" button. In our case, it is the Latin Square and it is therefore not necessary to search for another experimental design by optimization. The designs which have a distance of 0 are exactly adapted to the desired design. Click on the OK button to start the computations.Ī new dialog box pops up showing the proposed orthogonal designs coming from the internal database and which are closely related to the problem. In the Options tab, leave the default options. In the General tab, select the table of qualitative factors, the number of responses (1), the number of experiences (9) and the number of repetitions (10). The Screening designs dialog box pops up. Setting up the dialog box for generating a screening designĪfter launching XLSTAT, click the DOE button on the ribbon and select Screening Sesigns. The effects of the factors on the response variables are assessed using ANOVA and the charts of the means. We use the XLSTAT screening designs for the design generation for the XLSTAT ANOVA for the analysis part. Our aim then is to study the effect of each factor. Once the design was generated, the shooting distances were tracked down. These factors are: the length of the rubber band (A1, A2 and A3), the ball position (B1, B2 and B3), the pullback distance (C1, C2 and C3) and the shooting height (D1, D2 and D3). The goal here is to analyse 4 factors having an effect on the distance to which the projectile launched by the catapult is sent. The data come from a classic example of a catapult, which is frequently used in training on experimental design. Dataset for the analysis of a screening design This tutorial will help you design and analyze a screening design in Excel using the XLSTAT statistical add-on software.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |