What should we measure?

Goal: Design the table that records the rounds played of the game with three doors, and decide along the way which facts about a game are noted and in which form their values come.

How to use: The table starts with three empty columns and two empty rows. Each column is a variable, which you name at the top and, if you like, give a data type below. When you move the mouse over the table, a + appears on the right for another column and one below for another game. Every cell can be edited, and Enter jumps to the row below. A value that does not fit the chosen type gets a yellow outline but stays in place. Finally you save the table as CSV or as RDS, and below the buttons stands the R code that reads the file back. The table stays stored in this browser until you choose "Start over".

Each row is an observation, here one game played. Each column is a variable, a fact we record for every game, such as the door with the car or the decision between staying and switching. All values in a column share one data type, and the label in the head of the column shows it the way R shows it for a tibble: <chr> for text, <dbl> for decimal numbers, <int> for whole numbers, <lgl> for TRUE and FALSE, and <fct> for categories with fixed levels. Without a chosen type, the label shows in grey which type R would infer from the values entered.