DATA COLLECTION:
The stimuli and experimental tasks were designed in PsychoPy3. The monkeys were housed in pairs in a vivarium and divided to participate in the experimental sessions, both for training and testing. The experimental apparatus consisted of a 15-inch touch-sensitive monitor (Elo TouchSystems, Menlo Park, CA) and a food pellet reward delivery system (Med Associates, St. Albans, VT). The monitor was attached to the front of the macaque's home cage. The reward delivery system was placed so that the food pellets dropped directly in the food box, located below the monitor. Stimulus presentation, reward delivery and data collection were performed with PsychoPy3. 
 
DETAILS OF THE FILE FORMATS USED TO RECORD THE DATA: 
The data (the selected location in each trial) were automatically recorded by PsychoPy3 in .csv format. DETAILS OF SOFTWARE USED: PsychoPy3 was used to create the stimuli and the experimental tasks. Data analysis was conducted using R 4.0.3. The stats, car, emmeans, rcompanion, and pwr packages were used to perform analysis. 

A DESCRIPTION OF HOW WE PROCESSED YOUR DATA TO OBTAIN YOUR RESULTS (INCLUDING THE STATISTICAL TESTS PERFORMED):
Data analysis was conducted by R 4.0.3.
Binomial test was conducted on the average accuracy of each test against chance levels for all the tests. The effect size, Cohen h, was calculated using ES.h() with pwr package. 
Wilcoxon Rank Sum tests were conducted on the accuracy as a function of the target side after collapsing targets on the same side for all the tests. The effect size was calculated using wilcoxonR() with rcompanion package. 
Proportional tests were conducted to compare the use of spatial cues versus ordinal cues for the Conflict test. The effect size, Cohen h, was calculated using ES.h() with pwr package. 
Binomial GLMs were conducted on accuracy (1 = correct or 0 = incorrect) as a function of the target side (left or right) and the number of dots (2, 4, 6 or 10) for the Number transfer test. 
GLMs on the number of first choices as a function of position (1L, 2L, 3L, 3R, 2R, 1R) and target side (left vs. right) with a Poisson distribution was conducted for the Conflict test. 
Binomial GLMs were conducted on accuracy (1 or 0) as a function of target positions (2L, 3L, 4L, 4R, 3R, 2R) and experimental condition (continuous transfer test versus ten-dot number transfer test) for the Continuous transfer test. 
Likelihood-ratio Chi-square was calculated using Anova with car package. Post hoc comparisons with Bonferroni correction were conducted using emmeans with emmeans package. 

DETAILS OF ANY SCRIPTS OR CODES USED:
NumberTransfer.r, Conflict.r, ContinuousTransfer.r: the first part of the codes were used to calculate the descriptive statistics, creating tables and plots. Then, exact binomial test, Wilcoxon Rank Sum test, and GLMs were conducted as described above. 

DESCRIPTIONS OF ANY COLUMNS AND TABLES IN YOUR DATASET (INCLUDING ABBREVIATIONS USED):
In the following files: (df_NumberTransfer.csv and df_ContinuousTransfer.csv), columns' descriptions are: A: correct_dot, the target location (1L, 2L, 3L, 4L, 4R, 3R, 2R, 1R); B: side_correct(left/right), the side on which the target located; C: FirstTouch, the location that participants selected on the trial (1L, 2L, 3L, 4L, 5L, 5R 4R, 3R, 2R, 1R); D: expName, the experiment (2-dot, 4-dot, 6-dot, 10-dot in df_NumberTransfer.csv; line, dots in df_ContinuousTransfer.csv); E: participant (Hooke, Sally); F: date, the date and time when the experiment was conducting; G: ord_position, the order of the target position regardless of the side (1st, 2nd, 3rd, 4th; for example, if the target location was 1L and 1R, ord_position would be 1st); H: correct (1 = correct response, 0 = error response). 
In df_Conflict.csv, columns' descriptions are: A: correct_dot, the target location (3L, 3R); B: side_correct(left/right), the side on which the target located; C: FirstTouch, the location that participants selected on the trial (1L, 2L, 3L, 3R, 2R, 1R); D: expName, the experiment (conflict); E: participant (Hooke, Sally); F: date, the date and time when the experiment was conducting; G: Ordinal, the correct location when ordinal cue was used (3L, 3R); H: Spatial, the correct location when spatial cue was used (2L, 2R); H: used_cue (spatial = choosing 2L(2R) when the correct location was 3L(3R), ordinal = choosing 3L(3R) when the correct location was 3L(3R), Neither = choosing other error locations). 

LICENCES/RESTRICTIONS PLACED ON ACCESS TO THE DATASET: 
All published metadata are released under a CC0 licence.