MATLAB: Are the values I get for the mean using “anovan” and “multcompare” not accurate

anovanfullmeanmodelmultcompareStatistics and Machine Learning Toolbox

I am using "anovan" and "multcompare" to view the averages of my data, which is broken up into many different groups. However, the mean for some groups is higher than any individual value in that group, so the average seems to be wrong.

Why am I observing this behavior?

Best Answer

The mean is not the value you expect because the default model is a two-way interactions model, meaning that it is not simply using the data points from that group, but fitting a model that involves all points and using it to predict into that group (so, the mean is a 'predicted' mean based on more data than just those 3 points). A 'full' model (what you get if you add the name-value pair 'Model', 'full' to your "anovan" command) is a three-way interactions model, so you can expect it to reproduce the data if there are exactly three grouping variables.

For example, add the name-value pair to your "anovan" command by changing

>> [p,tbl,stats,terms]=anovan(data, groups);

>> [p,tbl,stats,terms]=anovan(data, groups, 'Model', 'full');

You can read more about the 'Model' name-value pair on the following documentation page:

https://www.mathworks.com/help/stats/anovan.html#bulw41w-model

Related Solutions

MATLAB: How to ensure that “stepwiselm” only removes terms from a model without ever adding any

For "stepwiselm"s default SSE criterion, a term is added when the SSE is smaller than 'PEnter' and is removed when it is larger than 'PRemove'. You can therefore make sure that "stepwiselm" never adds a term when using the SSE criterion by setting the value of 'PEnter' to 0 (or less). You may need to specify a different value of 'PEnter' if using some other value of 'Criterion' - please refer to the following link for documentation on the behavior of 'PEnter' and 'PRemove' for different criteria:

https://www.mathworks.com/help/stats/stepwiselm.html

You may, however, find that using the above method results in some desired terms not appearing in the computed model, since they may have been removed at some stage of the process without the ability to later add them back in. If this is the case, consider removing the PEnter = 0 property and instead use the 'Upper' input to "stepwiselm" to ensure that undesired terms are not added to the model. Details on how to supply the 'Upper' argument can also be found at the above link.

MATLAB: How to use custom date labels for the x-axis in MATLAB plots

It is more convenient to format the tick labels using the "XTick", "XTickLabelMode" and "XTickLabel" properties of the axis object (vs. using "datetick" function):

>> figure
>> box on
>>
>> % Start, end and number of ticks...
>> startDate = 7.3457e+05;
>> endDate = 7.3458e+05;
>> numbertick = ceil(endDate - startDate);
>>
>> % Plot them
>> xData = linspace(startDate, endDate, numbertick);
>> plot(xData,ones(size(xData)));
>>
>> %%Modify the axes properties
>> set(gca, 'XTick',xData)
>> set(gca, 'TickDir', 'out');
>> set(gca, 'XTickLabelMode', 'auto');
>>
>> % Modifying the labels
>> labels = get(gca, 'XTickLabel');
>>
>> for i = 1:length(labels)
>>     if i == 1 || i == ceil(length(labels)/2) || i == length(labels)
>>         labels{i} = datestr(xData(i), 'mm/dd');
>>     else
>>         labels{i} = '';
>>     end
>> end
>>
>> set(gca, 'XTickLabel', labels)

Best Answer

Related Solutions

MATLAB: How to ensure that “stepwiselm” only removes terms from a model without ever adding any

MATLAB: How to use custom date labels for the x-axis in MATLAB plots

Related Question