Two Way Anova with non replication

owenbyer

I am trying to run a Two-Way Anova test with two factors (three levels of cylinders and two levels of transmission), and I have one data value in each block (so 6 total).  This means I have no replication, and Excel handles it just fine, because there is an option to not have replication. In Jasp, however, I get an error message:

https://mail.google.com/mail/u/0?ui=2&ik=0e3f9c8ef1&attid=0.1&permmsgid=msg-a:r-7785051184714967187&th=18048682e32aae82&view=fimg&fur=ip&sz=s0-l75-ft&attbid=ANGjdJ_ppaKrnYb8SvAbjMZlKSNEkEhbo3I_WorqsBPFjzwGiNurfYk76r5iYMkr65djVoBq7YVobc-9Xp3pb8UG6YVND6qSp9aNx9S04x2-oUqDv4S03vkcXwcc0ig&disp=emb&realattid=ii_l27yf9wc0

Jasp works fine for two way Anova when there are at least two data values in each block.  Is there some way to get Jasp to handle the nonreplication scenario?

andersony3k

I think that if Excel handles it just fine, Excel is doing it wrong. For an ANOVA, you need to estimate variance across conditions as well as error variance (variance between conditions). An ANOVA can be calculated only if at least one of the six cells contains more than one observation, thus permitting a within-condition variance to be calculated.

[The so-called "two-factor" ANOVA without replication" is a misnomer in Excel, it's actually a one-factor repeated-measures ANOVA (maybe Excel is trying to call the "participant" variable a "factor"). But that's not the kind of data you have, so Excel won't analyze it. See https://www.dummies.com/article/technology/software/microsoft-products/excel/use-anova-two-factor-without-replication-data-analysis-tool-excel-227408/ (I'm not name-calling, here. That's just the title of the website.)]