When can Prism cope with missing values in two-way ANOVA?

Whether or not Prism can handle missing values depends on how you enter the data and whether you are asking for repeated measures ANOVA.

No repeated measures with raw data entered

Prism can handle missing values in two-way ANOVA without repeated measures so long as there is at least one value for every row/dataset. 

In the table below, one replicate is left blank (is missing). Prism analyzes these data just fine. Prism uses the method detailed by Glantz and Slinker (1) reporting Type III sum-of-squares.

In the table below, there are no values entered for the second row, data set B. Prism cannot analyze these data.

 

No repeated measures with data entered as mean, n and SD or SEM

f your data are balanced (same sample size for each condition), you'll get the same results if you enter raw data, or if you enter mean, SD (or SEM), and n. If your data are unbalanced (different n for different cells), it is impossible to calculate precise results from data entered as mean, SD (or SEM), and N. Instead, Prism uses a simpler method called analysis of “unweighted means”. This method is detailed in LD Fisher and G vanBelle (2).   If your data are almost balanced (just one or a few missing values), the approximation is a good one. When data are unbalanced, enter individual replicates (not mean, SD/SEM and n) whenever possible.

Repeated Measures in one factor

Prism can compute repeated measures two-way ANOVA fine if alternative treatments were given to different numbers of subjects.

Prism cannot compute repeated measures two-way ANOVA if values at some time points are missing.

For example, assume that the matched values are stacked into a subcolumn, which is the most common way to enter data.

It would be fine if column A had data in three subcolumns, while column B had data in only two subcolumns. You need to format each data set with the same number of subcolumns, but could simply leave a subcolumns blank for data set B. Prism would perform repeated measures ANOVA fine in this case.

Prism cannot perform repeated measures ANOVA is, say, the value in row 2 of subcolumn 2 of data set B was missing. Computing ANOVA with missing values is far from straightforward, and Prism doesn't attempt it.  

Repeated Measures in both factors

Prism cannot compute repeated measures two-way ANOVA with repeated measures in both factors if there are any missing values.

1. SA Glantz and BK Slinker, Primer of Applied Regression and Analysis of Variance, McGraw-Hill, 1990.

2. LD Fisher and G vanBelle, Biostatistics, John Wiley, 1993