OK, I'll play.  One of the first things I think about is measurement error.  I don't trust any measurement system unless it's been studied. Typically, I'll start with sampling to understand the stability and size of the measurement errors (precision: repeatability & reproducibility) and compare those sources to the other sources I wish to discriminate. ... View more

I'll make this my last response as well. All I can say is I tried. I'm sorry I am having trouble communicating my points. Here are links to reports that tweets from @realDonaldTrump may not be from just the president. https://www.npr.org/2018/04/19/603904858/who-is-the-mystery-man-behind-realdonaldtrump-besides-the-president https://www.wired.com/story/tell-when-someone-else-tweets-from-realdonaldtrump/ https://www.wnycstudios.org/podcasts/takeaway/segments/president-trumps-tweet-co-pilot Google it. If you can't trust the measurement system I'm not sure what analysis you have confidence in?   I already gave you multiple sources of the correct use and interpretation of control chart method. And yes there are many, many examples of the inappropriate use of control charts.  This quote is from your reference " Have you heard that data must be normally distributed before you can plot the data using a control chart?  Quite often you hear this when talking about an individuals control chart.  This is a myth.  Data do not have to be normally distributed before a control chart can be used – including the individuals control chart. ".   Deming's "Out of the  Crisis" (https://mitpress.mit.edu/books/out-crisis) Deming's paper: https://www.jstor.org/stable/i326388 Shewhart's book: https://books.google.com/books?id=EoynRAl0Po4C&source=gbs_similarbooks   By the way, it looks like you only used data from 2020.  Here is a graph of tweets since 2009. It also has multiple categories for Original, re-tweet, Replies, Links/images, Tweets.   ... View more

Echoing @P_Bartell there are many options to model data. Regarding #2, I don't understand what your goal is?  It sounds like you want to create >1st order polynomials.  Is this from observational data or from an experiment?  I don't understand what you mean when you say " I will call the platform in JMP with a BY variable " ? What platform? Not to be  belligerent, but you do realize the reason linear regression is called linear is because the coefficients are linear, not the terms in the model. ... View more

You can actually still do it by changing all of the variables to continuous just for this platform. I'm not interested in the statistics per se, but the plots. Agreed the categorical variables will look a bit strange, but obvious relationships can show up. For example, in the image I have attached I have categorical factors of Shift and Technician with an ordinal Y.  You can see, obviously, the correlation of shift and tech.   ... View more

@ryan_lekivetz, sorry I apparently did a poor job of explain what we discussed yesterday. Thanks for your clarification. ... View more

I do have wonder what it is you are working on where you suspect cubic effects? Even if they exist, they will be difficult to "manage". Seems you might want to investigate alternative repossess variables that may be robust to those type of effects? ... View more

AFAIK, in the custom design platform, as long as you select Necessary (vs. If Possible), JMP will create a design that will allow for estimation of those effects. ... View more

I spoke to a JMP DOE developer yesterday (JMP summit) and he confirmed the plots would always show the correlation, but in actuality with the correct design, there is no correlation. ... View more

As a start, use the Analyze>Multivariate Methods>Multivariate platform.  Put all of the variables in the Y, Columns window.  You'll get correlation coefficients and more importantly a scatter plot matrix to look for collinear relationships. ... View more

I'm sorry, my understanding is that you will always get that correlation due to the coding that goes on "behind the scenes".  What I meant to say, AFAIK, the x and x^3 will actually not be correlated if the appropriate designs selected and the actual values are used.  The correlation matrix will always show the correlation due to the coding. ... View more

Ella,   What you are seeing is a result of the coding of values for the factors.  -1^3 = -1, 1^3 = 1   So if you use "actual values" you will see the correlation matrix without those terms correlated. ... View more

Assuming you have 4 factors at 4 levels, when using the Customer Design platform, in the Model, highlight the effects you want. If it says If Possible, you can change that to Necessary.     ... View more

I don't know what version of JMP you are using?  I'm also not sure which platform you want to remove the lines from, but typically on the plot, if you right click anywhere on the graph, go to Customize and there are a list of items you can add or "remove" or change the color or line width for.      ... View more

Again, I'm not sure I understand the question?  If it is about learning how to use residuals to determine adequacy of your model, there are many books and papers on the subject.  First, a residual is the difference between the prediction from the model and the actual value.  Regardless of any assumptions required or implied, in order for the model to be useful, you would hope that it does a fairly good job of predicting the actual results.  From a model development perspective: 1. you would hope the model isn't biased high or low (the residuals would be distributed around 0), 2. you would hope the residuals have about the same variation around the model and this variation isn't fluctuating greatly ( the residuals have a constant variance) 3. you would hope there aren't any unusual data points not explained by the model (absence of outliers in the residuals) 4. You would hope the residuals didn't form some pattern or were related to each other (independently distributed) If some of these hopes are not satisfied, you should seek to understand why.  You should challenge the effectiveness of your model and how you can modify it to make it more useful (and better yet arrive at a better understanding of what is actually going on). ... View more

I'm a bit confused by this " Our understanding is that to test the assumption of normality in ANOVA".  There is NO assumption of  normality of the raw data analyzed via ANOVA.  The only assumptions are NID(0, variance).  That is normally and independently distributed residuals with a mean of 0 and a constant variance.  There are a number of ways to look at residuals: 1. in a time series, 2. vs. predicted 3. as a distribution, 4. leverage plots There are a number of types of residuals as you have listed some.  There is no right one to use.  Easy enough to look at them all. ... View more

https://www.jmp.com/support/help/en/15.2/?os=mac&source=application&utm_source=helpmenu&utm_medium=application#page/jmp/interactive-effect-summary.shtml   ... View more

In general, test of significance (p-values) are tests based on rejecting or not rejecting the Null Hypothesis.  The p-value is the associated probability of a Type 1 or alpha error (rejecting the Null when the Null is true). These tests can be done on different statistics with different associated probability distributions (e.g., t, F, Chi square, etc).  The Parameter Estimates reports the coefficients of the model and Prob>|t| is the probability a value greater than the calculated t Ratio (Estimate/Std Error) could exist by chance alone (this is actually area of the probability distribution, in this case Student's t-distribution beyond the calculated test statistic) .  Effects tests use the calculated F ratio (mean square effect/mean square error) in the same way.  The Prob>F is the p robability a value greater than the calculated F-Ratio could exist by chance alone (i.e., when the effect = 0).  This picture might clarify graphically what is being calculated.   ... View more

Roly, Always happy to help folks "broaden their horizons"! Perhaps a bit more broadening...Your statement " now understand that the distribution does not have to be normal, but some adjustment is needed if it is non-normal." is incorrect and highlights a major misconception of the control chart methodology (if you disagree, perhaps you can site the section of the literature that supports your argument).  If the problem were enumerative, I would be interested in the distribution (and the probabilities associated with it), but since the problem is analytical, the distribution is of little consequence. I suggest you read  Deming "On Probability As a Basis For Action", The American Statistician, November, 1975, Vol. 29, No. 4.  Admittedly, it would appear more rational if the lower control limit were truncated to 0, but most if not all users recognize this in their assessment of the charts.  The software is simply doing the math without regard to interpretation.  I would say most if not all statistical outputs need to be interpreted IN CONTEXT!    Regarding your comment: "Sorry for the confusion about consistent. I am not quite sure what you mean by "The # of tweets is inconsistent". If it is that the points are not distributed normally, I agree, they appear to be distributed with a gamma distribution."  I understand this concept is difficult to understand.  Perhaps this is the reason so many do not know how to use control charts appropriately. I'll say this again, inconsistency has nothing to do with a distribution. p. 359, Deming, "Out of the Crisis".  The beauty of the Shewhart control charts is they assume no distribution.  They provide a graphical look at the data in a time series (this is lost in any distributional analysis!).  "A distribution only presents accumulated history of performance of a process, nothing about its capability...The capability of a process can be achieved and confirmed by use of a control chart, not by a distribution", Deming, "Out of the Crisis, P. 314. The range chart answers the question: Is the variation within subgroup (due to the variables changing within subgroup) consistent, stable and therefore predictable?  Consistent in this reference means within an expected, predictable amount of variation.  For the MR chart, the question being answered is similar: Is the variation in consecutive data points (due to the variables changing between consecutive data points) consistent, stable and therefore predictable?  So when I say the number of tweets is inconsistent, this means the process of posting tweets is not predictable, not random and unstable.  It would be difficult to summarize, with any confidence, this process with any enumerative statistics or make any probability statements as the statistics/probabilites would depend on when you looked at the data.   Regarding the data source...You can't possibly answer who sent the tweets, only that they came for some account registered to "Donald".  I do not have direct knowledge of the actual number of tweets, nor who sent them.  Is there bias in the measurement process? Are the numbers accurate?  I can't answer that.  But I could suggest a sampling plan to provide insight.   The distributional analysis with associated probabilities are inappropriate for a process that is inconsistent, non-random and unpredictable.   I leave you with this quote, again from Deming (paper referenced above, : “Analysis of variance, t-test, confidence intervals, and other statistical techniques taught in the books, however interesting, are inappropriate because they provide no basis for prediction and because they bury the information contained in the order of production. Most if not all computer packages for analysis of data, as they are called, provide flagrant examples of inefficiency.” ... View more

On the left column you will see a box with all of the columns listed.  if you select the colored geometrical shape next to the variable, you can see the different modeling types available.  Your Y is currently Nominal (red histogram shape).  You will need the data type to be continuous to run the platform you want.  If you can't change it in the left window, you will need to highlight the Y column, right click and choose the dropdown window to Column Info.  There you can change Data type and modeling type. ... View more

Welcome to the JMP community.  It would be helpful in the future that you let us know what version of JMP you are using and which OS.   The simple answer is YES.  I'm not sure what you mean by # of measurement data?  Are these from different locations on the experimental unit?  Why do you need 1000 measures of friction force???? Even though all of the data is taken (it seems automatically), can't you decide which values to use to evaluate the experiment?  I would always suggest you have hypotheses to explain why you are collecting the data you are collecting! (Because we always do it that way is not acceptable). If you can't reduce the # of data points for each of the 4 Y's, I would collect the data and create a column for every data point for each treatment combination (yes, 1080 columns: B1-B20, H1-H10, C1-C50, F1-F1000).  From there, you can remove columns that aren't informative (don't vary enough to be practically significant), you can run Multivariate Methods> Multivariate to reduce the columns that are strongly correlated. Then for each Y (B, H, C, F) stack those columns.  You can look at the variability within treatment (variability charts or S-charts). Then, if appropriate, summarize the stacked columns (e.g., mean and variation) by the treatment combinations.  Then proceed with analysis for each of those summarized Y's. ... View more