| Due Date | Topic | Assigned to |
|---|---|---|
| 11.06.19 | Familiarize with topic | all |
| 11.06.19 | Hypothesis and structure of Experiment | all |
| 13.06.19 | Rücksprache mit Franke | all |
| 26.06.19 | Translation of prompts | Caro |
| 26.06.19 | Instructions | Caro |
| 26.06.19 | Experiment coding | Marc, Lukas |
| 26.06.19 | First trial of experiments testing for usability/clearness | all |
| 26.06.19 | Coding of statistical analysis | Polina |
| 28.06.19 | Pre-registration report | all |
| tbd | Data acquisition | all |
| tbd | Statistical analysis | Polina |
| tbd | Writing of report | |
| 19.07.19 | Expected submission date | --- |
| Task | Assigned to | Due date | Done? |
|---|---|---|---|
| write welcome text | Caro | 20.06.19 | ✅ |
| translate questions | Caro | 20.06.19 | ✅ |
| change text from "Dein/e/er" zu "Der/Die/Das" in prompt generator | Marc | --- | ✅ |
| change sentences in polgrice... | Caro | 21.06.2019 | ✅ |
- Hypothesis formulation and structure of Experiment
- Two experiments or two parts of one?
- Stimuli prep
- experiment 2
- Statistical Analysis
- Which scale would be best to replicate for second experiment? (hearts, numbers, starts, ...)
- experiments in one or separate? only one experiment (experiment 2)
- Scale?
- Übernehmen von Herzen
- ordinale Regression
- adjustment of names ?
- randomizing names
- due date? ✅
- hypotheses ok?
- can mean be left out?
- General conditions
- 5 states (1-5 hearts), uniform prior distribution
- utterances {terrible, bad, ok, good, amazing}
- 3 goals: honest, nice, mean
- Exp 2: measurement of listener's inference about state given utterance and goal
- 35 participants, context items with question about performance, 15 randomized items / participant (3 goals x 5 states), rating task on heart scale
- hypothesis 1: positive utterances:
- mean(nice_inferred) < mean (honest_inferred)
- mean(nice_inferred) > mean(mean_inferred)
- hypothesis 2: negative utterances:
- mean(mean_inferred) > mean(honest_inferred)
- mean(honest_inferred) = mean(nice_inferred)
- when coding data analysis, ordinal scale brm will be used (group ontervals are fitted)
- Sequence
- Welcome view
- Task description
- 15 randomized trials
- same gender per task