Wir haben folgendes Design:
- AV: Akzeptanz von Algotithmen
- Es gibt drei Gruppen: a) ohne Hinweistexte, b) mit sachlichen Hinweistexten c) mit emotionalen Hinweistexten
- Die Probanden sollen die Akzeptanz von Algorithmen bei insgesamt zwanzig verschiedenen Szenarien angeben. Diese Szenarien haben jeweils eine Bewertung auf den Skalen Risiko (1=niedrig .. 5=hoch) und Kreativität (1=niedrig .. 5=hoch)
Gewünscht ist eine Auswertung, wie sich die Hinweistexte auf die Akzeptanz auswirken unter der Moderation von Risiko und Kreativität des Szenarios.
Der Faktor Hinweistext ist offenbar between subject mit drei Ausprägungen. Aber wie gehe ich mit den Faktoren Risiko und Kreativität um? Ist das eine Messwiederholung (ANOVA)? Wie berechet sich die benötigte Stichprobengröße?
Ich würde mich über sachdienliche Hinweise freuen.
This sounds like a mixed-design experiment, with the factor of "hint text" being a between-subject factor and the factors of "risk" and "creativity" being within-subject factors.
To analyze this design, you would likely use a mixed-design ANOVA (analysis of variance), also known as a split-plot ANOVA. This allows you to examine both between-subject and within-subject effects simultaneously.
The between-subject factor is "hint text" with three levels: without hint texts, with factual hint texts, and with emotional hint texts. The within-subject factors are "risk" and "creativity," each with five levels corresponding to the rating scales (1=low to 5=high).
To calculate the required sample size, you would need to consider factors such as the desired power of the study, the expected effect size, and the significance level. You can use statistical software or online calculators specifically designed for sample size calculation in ANOVA.
Given the complexity of the design and analysis, consulting with a statistician or someone experienced in experimental design and analysis would be beneficial to ensure appropriate methodology and sample size determination.
Achim Scheuss , you should rather thank ChatGPT. Chuck's answer has a 100% chance of being AI generated (according to https://www.scribbr.com/ai-detector/).
Further note that your question cannot be seriously answered until you describe in detail what your DV and your sample size is so that sufficient information is available to judge a sensible distributional model or if one may possibly rely on the central limit theorem.
An additional aspect not at all recognized by the AI is that your two predictors are ordinal. I think this is relevant and should be considered in the analysis. I don't know how to do this with ordinal data, but maybe it is already helpful to not analyze the data but to analyze their ranks instead and use a multiple regression model including an interaction term. In any case I'd suggest to think about a good visualization of the results like in a heatmap (2 dimensions for the predictors, DV value is color-coded) or something similar to recognize and interpret general patterns.
A serious statistical analysis will surely include extensive simulations that would need to be programmed. This will require a consultation of a statistician being experienced in such things. Here I do agree with the advice given by the AI.
PS: just for comparison, this text has a 5% chance of being generated by AI ;)
Achim Scheuss can you say more about the dependent variable? How is it measured?
I already wrote an answer, but then I realized that you only mentioned the (ordinal) scales of your covariables "Risiko" and "Kreativität", but you did not say anything about the DV itself. Therefore, I deleted it.
And maybe you could add an english version of your question, so that all users are able to profit from the discussion here?
- Badges
- Science topic
- Similar topics
- Engineering
- Petroleum Engineering
- Oil