https://doi.org/10.5061/dryad.fqz612jzp

Description of the data and file structure

See descriptions of the datasets provided in the order that they appear in the manuscript.

IGI = intergroup interaction (see manuscript).

Life-history datasets

Datasets which were used to investigate the reproductive consequences of intergroup conflict using long-term life-history data from the Dwarf Mongoose Research Project.

Breeding_attempts_binomial.csv - dataset used to analyse whether IGI rate influenced the number of breeding attempts. Response variable MORE_2_BR_ATTEMPTS (Yes/No), refers to was there more than two breeding attempts in the season. Each row is a breeding attempt by a group. Predictor variables defined below. 

Interlitter_interval.csv - dataset used to analyse whether IGI rate influenced the period between the birth of the first and second litter (number of days between the first and second litter). Response variable referred to as INTER_LITTER_INTERVAL in the dataset. Each row is an inter-litter interval. Predictor variables defined below. 

Pup_survival_overall_poisson.csv - dataset used to analyse whether IGI rate influenced the number of pups alive at 3 months post-emergence. Response variable referred to as TOTAL_PUPS_SURVIVE_3M in the dataset. Each row is a litter by a group. Predictor variables defined below. 

Pup_emergence_poisson.csv - dataset used to analyse whether IGI rate influenced the number of pups to emerge from the breeding burrow. Response variable referred to as TOTAL_PUPS_EMERGE in the dataset. Each row is a litter by a group. Predictor variables defined below. 

Pup_survival_post_emergence.csv - dataset used to analyse whether IGI rate influenced the proportion of pups surviving post-emergence. Response variable binds two columns referred to as PUPS_SURVIVED_3M and PUPS_DIED_3M. Each row is a litter by a group. Predictor variables defined below. 

IGI_poisson.csv - dataset used to analyse whether group size and/or adult body mass influenced the number of IGIs a group had. Response variable referred to as IGI TOTAL. Each row is a litter by a group. Predictor variables defined below. Model included an offset (log number of observation sessions) to account for differences in the number of observation sessions (SESSION_TOTAL in .csv file). 

Variables used in the life-history analyses

Where appropriate, we also included the following predictor variables. For relevant time periods see the manuscript.

• IGI_RATE = the rate of intergroup interactions calculated as the number of IGIs in a relevant timeframe (IGI_NUMBER) divided by the number of observation sessions (SESSION_NO) in the same timeframe. An observation session refers to a 4 hour period where an observer was with a mongoose group (either morning or afternoon). 
• WEIGHTED_ADULT_GRP_SIZE = calculated to account for the varying number of adult individuals in a group over a given period. For instance, if in a time period of 90 days, group size was 8 adults for 30 days, 9 adults for 45 days and 10 adults for 15 days, the weighted group size is 8.83 adults, calculated as ((8x30)+(9x45)+(10x15))/90. 
• RAINFALL = the total amount of rainfall recorded in a relevant timeframe (ml).
• AVERAGE_MORNING_WEIGHT_ALL = mean adult body mass, calculated from all morning body-mass measurements from group members during the relevant period (grams).
• LOSE_1ST_2ND_LITTER = whether the dominant female lost her first and/or second litter, Yes or No.

• DOM_F_CHANGE = whether there was a changeover in the dominant female, Yes or No.
• FIRST_TIME_DOM_BR = whether it was the first time that the dominant pair had bred together, Yes or No.
• SUB_F_BIRTH = whether a subordinate female had given birth at the same time as the dominant female, Yes or No.
• LITTER_NO = the breeding attempt number in the season (first, second, third).

Sentinel datasets

We conducted observations of sentinel behaviour following both IGIs and latrining events to compare with observations in matched-control periods. For the post-IGI analyses, the observation periods are referred to as IGI and Control, with each period 60 min in duration. For the post-latrine analyses, the observation periods are referred to as Latrine and Control, with each period 30 min in duration. 

Sentinel_post_IGI.txt - dataset used to analyse the proportion, rate and bout duration of sentinel behaviour after IGIs, along with the proportion of adult group members involved in sentinel behaviour. 

Sentinel_post_latrine.txt - dataset used to analyse the proportion, rate and bout duration of sentinel behaviour after latrines, along with the proportion of adult group members involved in sentinel behaviour. 

Variables used in the sentinel analyses

• Proportion_sentinel_present = proportion of time that individuals acted as a sentinel in the observation periods. Calculated as the total time sentinels were present (min) divided by the duration of the observation session (min). 
• Sentinel_occurence_rate  = total number of sentinel bouts divided by the duration of the observation session. 
• Proportion_of_group_involved = the number of adult individuals that conducted sentinel behaviour during the observation period divided by the total number of adults in the group. 
• Mean_scan_duration = mean sentinel bout duration (seconds) during the observation period.