@[toc](Sections)
### Basics
Great, now we know how to collect data. But how do you access data? Ibex simply generates a file called *results* under the project's folder named... *results*. It is a simple CSV file (with lots of comment lines, starting with `#`) which makes it very easy to load in R.
### Loading the *results* file in R
```R
# To summarize the data (see %>% commands below)
require("dplyr")
# The fields are separated by commas
# There are comment lines starting with # (useless for analyses)
# There is no header (no single line assigning each column a name)
results_full <- read.csv("results", sep=",", comment.char="#", header=F)
# As we can see, the columns have default Vn names
head(results_full)
# The 7 first columns (until IbexGroup) are not particularly relevant
# The 8 last columns (starting from Parameter) contain all we need for our analyses
colnames(results_full) <- c("SendResultsTime", "MD5IP",
"IbexController", "IbexItemNum", "IbexElementNum", "IbexLabel", "IbexGroup",
"Parameter", "Value", "Time", "Comment", "ID", "Label", "Item", "Group")
```
### Subsetting to trials
```python
# Parameter == "answer" means a response line for the element we labeled 'answer' (our scale)
results_answer_ratings <- results_full %>%
filter(Label == "rating" & Parameter == "answer")
# Parameter == "alternative" means a response line for the element we labeled 'alternative' (our input box)
results_answer_inputs <- results_full %>%
filter(Label == "input" & Parameter == "alternative")
```
### Taking a quick look at the data
```python
# Let's take a look at all the alternatives that were proposed for each item
results_answer_inputs %>%
group_by(Item, Value) %>%
summarize(
n = n()
)
# Let's see how 'any' vs 'either' impacted perception of naturalness
results_answer_ratings %>%
group_by(Group) %>%
summarise(
n = n(),
score = mean( as.numeric(as.character(Value)) )
)
# Let's look at whether 'any' vs 'either' influenced rate of antonymic alternatives
results_answer_inputs %>%
mutate(dry = grepl("dry", Value), cold = grepl("cold", Value)) %>%
group_by(Group, Item) %>%
summarise(
dry = sum(dry),
cold = sum(cold)
)
```
### Comparing two lines (= two measures) for the same trial
```python
# Were people faster to score in the 'any' or the 'either' group?
results_RT_per_item_per_ppt <- results_full %>%
filter(Label == "rating" & (Parameter == "answer" | Value == "RunFirstInstruction" )) %>%
group_by(Item, ID) %>%
summarize( Group = unique(Group), difference = diff(as.numeric(as.character(Time))) )
results_RT_per_group <- results_RT_per_item_per_ppt %>%
group_by(Group) %>%
summarize( meanRT = mean(difference) )
```
### Saving simplified table for manual edits
```R
# Let's save a smaller table so we can manually treat the alternatives in a spreadsheet editor
results_simplified <- results_full %>%
select(Parameter:Group) %>%
filter(Parameter != "Page")
write.csv(results_simplified, "results_simplified.csv")
```
[1]: https://osf.io/9trsx/wiki/Recruitment%20&%20Data%20collection/
[2]: https://osf.io/9trsx/wiki/Preloading%20Resources/
