Maternal Stress and Food Preferences - Evidence from Essex

Last registered on October 27, 2020

Pre-Trial

Trial Information

General Information

Title
Maternal Stress and Food Preferences - Evidence from Essex
RCT ID
AEARCTR-0003410
Initial registration date
October 13, 2018

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
October 19, 2018, 6:20 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated
October 27, 2020, 4:22 PM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Primary Investigator

Affiliation
University of Bristol

Other Primary Investigator(s)

PI Affiliation
European University Institute
PI Affiliation
University of Bath
PI Affiliation
European University Institute

Additional Trial Information

Status
Completed
Start date
2018-10-15
End date
2019-03-31
Secondary IDs
Abstract
The spread of obesity in developed nations over the past decades has disproportionally affected lower socioeconomic groups as they face a larger risk of overweight and obesity (McLaren, 2007). Among the factors suggested to cause this socioeconomic gradient in obesity are higher levels of stress and a lack of tools to cope with stressful situations among low socioeconomic groups (Moore and Cunningham, 2012). Eating more or eating foods high in fat and sugar might provide comfort in stressful times. Making healthy decisions may also be more difficult when the mind is occupied with handling stressful situations.

This study focuses on two main research questions:
(1) How does acute stress affect food choices in the context of immediate and planned consumption?
(2) How does in utero exposure to maternal stress affect the child’s future preferences for obesogenic and non-obesogenic foods?

To study these research questions we propose to conduct a lab experiment with approximately 200 low income mothers. Experimental sessions will be pre-assigned to a stress treatment group or to a control group. In the stress treatment, subjects will be asked to complete a 10 minute block of short incentivised decision tasks which is designed to mimic stressors often experienced by low-socioeconomic mothers: making decisions with consequences for others subject to financial and time constraints as well as distractions.

Following the stress or control task, subjects will be asked to use a fixed budget to purchase food items in a “virtual supermarket”, a computer-based tool similar to online supermarkets. For this food shopping task, sessions will be further pre-assigned to one of two food shopping environment treatments: a simple or a complex choice environment.

During a five minute break and during a questionnaire on demographics, participants are furthermore offered a choice of snacks involving high- and low-calorie snack foods. The nutritional content of the chosen food shopping basket and the quantity of snacks consumed will be used to determine the impact of acute stress on immediate and planned food consumption choices.

In addition to the randomly assigned acute stress and shopping environment treatments, our study will aim to capture chronic stress during pregnancy using a questionnaire on specific stressors and their perceived stressfulness. This measure will be used to estimate the impact of chronic stress during pregnancy on questionnaire-based measures of children’s food preferences.

We previously conducted a study with a similar design in Italy (AEARCTR-0003089).
External Link(s)

Registration Citation

Citation
Belot, Michèle et al. 2020. "Maternal Stress and Food Preferences - Evidence from Essex." AEA RCT Registry. October 27. https://doi.org/10.1257/rct.3410-2.0
Former Citation
Belot, Michèle et al. 2020. "Maternal Stress and Food Preferences - Evidence from Essex." AEA RCT Registry. October 27. https://www.socialscienceregistry.org/trials/3410/history/78621
Experimental Details

Interventions

Intervention(s)
Intervention Start Date
2018-10-15
Intervention End Date
2018-10-19

Primary Outcomes

Primary Outcomes (end points)
i) Food shopping choice during the experimental session: calorie (kcal), saturated fat (g) and sugar content (g) of the chosen basket

ii) Snack intake during experimental session: consumption of low-calorie snack (in grams), consumption of high-calorie snack (in grams)

iii) Food preferences of youngest child:
- Mediterranean Food Preference Index based on liking of food groups in KIDMED Mediterranean Diet Quality Index for children and adolescents (Serra-Majem et al, 2004)
- BMI percentile (age & gender specific)
- Food consumption: KIDMED Mediterranean Diet Quality Index for children and adolescents (Serra-Majem et al, 2004) excluding “commercially baked goods / pastries for breakfast” item
Primary Outcomes (explanation)
Mediterranean Food Preference Index is calculated using the following food groups and respective index weights: fruit (+2), vegetables (+2), legumes / pulses (+1), fish / shellfish (+1), pasta / rice (+1), cereal / grains (+1), nuts (+1), dairy products (+1), commercially baked goods / pastries (-1), sweets / candies (-1), fast food (-1), sugar-sweetened drinks (-1, additional non-KIDMED item). The average liking of foods in each food group (based on 5-point Likert scale) are summed using the above weights to produce the index.

Secondary Outcomes

Secondary Outcomes (end points)
i) Food shopping choice during the experimental session:weight of chosen fruit and vegetables (g)

ii) Snack intake during experimental session: total calorie (kcal), saturated fat (g) and sugar content (g) of the consumed snacks (deterministic functions of the two primary snack outcomes)

iii) Food preferences of youngest child: Taste preferences
Secondary Outcomes (explanation)
The taste preferences are measured as the average liking of sour, salty, umami, bitter and sweet foods respectively. The following foods are used for the tastes:
- Sour: grapes, oranges, lemons, balsamic vinegar
- Salty: chips / French fries, crisps
- Umami: mushrooms, parmesan, meat & fish foods
- Bitter: asparagus, Brussels sprouts, black olives
- Sweet: pastries, cake, biscuits, chocolate, ice cream, cola, Fanta

Experimental Design

Experimental Design
The sessions for this experiment will take place between 15 October and 19 October 2018 in the experimental laboratory of the University of Essex, UK. Sessions will last approximately two hours and will start at 10:30 am, 2:00 pm or at 5:00 pm. We employ a between subject design, with subjects only attending one session. The session slots have been pre-assigned to the four experimental conditions.
Experimental Design Details
The experiment will follow a 2x2 experimental design resulting in 4 experimental conditions: 1) Stress Task & Simple Shopping Choice, 2) Stress Task & Complex Shopping Choice, 3) Control Task & Simple Shopping Choice, 4) Control Task & Complex Shopping Choice.

These experimental conditions will be pre-assigned at the session level. This pre-assignment of sessions to experimental conditions ensures balance in terms of day of the week and time of day. When signing up for participation in the experiment, participants are asked to indicate their preferred session slots, but will not be informed in advance of the treatments associated with each time slot. If participants indicate availability for multiple slots, they are assigned to one of the slots solely based on scheduling concerns.

Upon arrival at our lab facilities, participants’ body weight and body height is measured without shoes and heavy clothing. Throughout the experimental session, participants are asked to wear an armband monitoring their heart rate using an optical sensor. At the beginning of the experimental session, participants are asked to provide a first saliva sample.

Following this, participants are asked to complete a 10 minute task. The nature of the task depends on the session’s randomly assigned experimental condition:

In conditions 1) and 2) (detailed above), i.e. the stress treatment groups, participants are asked to complete an incentivised task aimed at inducing mild stress. More specifically, subjects are asked to complete a 10 minute block of short incentivised decision tasks. While the tasks are completed individually, incentives are based on the joint performance of “social groups”, each consisting of 2 participants in the same session which are randomly matched , to elicit social stress. Incorrect answers and incomplete tasks are penalized. Time pressure is induced by the tight overall time limit as well as by time penalties and time limits per task. Short incentivised knowledge questions will appear on screen at seemingly random times throughout the course of the task block to induce additional stress through distractions. This stress protocol is designed to mimic stressors often experienced by low-socioeconomic mothers: making decisions with consequences for others (e.g. for the family) subject to financial and time constraints as well as distractions (e.g. by children requiring attention).

In conditions 3) and 4), i.e. the control groups, participants are asked to complete a task of similar nature but with no stress inducing features. Specifically, they are asked to answer 14 simple knowledge questions after reading 7 short texts about a variety of topics. The correct answers to each question can be found in the corresponding text. The questions are similar to those asked as distractions during the stress task. Subjects are given 10 minutes for this task, there are no consequences from not completing all questions. The task is not incentivised and no “social groups” are formed. Asking simple questions after providing the answers guarantees that participants will not experience stress due to task difficulty. Due to the lack of individual or group incentives, participants should not feel under time pressure or social stress. Furthermore, there are no interruptions creating stress.

Following the first task, participants are asked to complete a “food shopping” task. Participants are given 10 minutes to allocate a fixed budget of £30 to food and drink items offered in the “virtual supermarket” tool. A variety of low-calorie and high-calorie food and drink items is available to choose from with prices matching market prices at a local supermarket. In total, approximately 150 products are on offer. Participants are encouraged to make their shopping choices as they would during a weekly shop at their local supermarket. The supermarket choice is incentivised: 1 out of 15 participants are randomly chosen to receive their chosen basket delivered to their home approximately two weeks after the session. This incentive scheme was chosen to motivate participants to make choices representative of normal shopping behaviour.

Participants are informed that if they are selected and have not spent the entire £30 budget, they will be paid the difference in cash up to £2 maximum. This is to discourage non-representative shopping choices aimed at spending exactly £30, and to ensure that participants do not feel the pressure to spend the exact amount, which could induce stress for all participants. Under this incentive scheme it is optimal for participants to aim to spend between £28 and £30.

The complexity of the food shopping environment depends on the experimental condition assigned to the session: In conditions 1) and 3) (detailed above), i.e. the simple shopping choice, products are listed separately in 10 different food categories. In conditions 2) and 4), i.e. the complex shopping choice, products are shown in a single long list, grouped by category but without labelling of categories. In both choice environments approximately 150 items are on offer. The order in which items are displayed within each category is randomized at the participant level to avoid order effects. Furthermore, the display order of categories and the first category shown when opening the supermarket tool are randomized.


After the “food shopping” task, participants are asked to provide a second saliva sample (approx. 25-30 mins after start of stress / control task) and then given a five minute break. After the break, participants are asked to complete a questionnaire on demographics, family characteristics and behaviours which might impact cortisol levels. During the break and the time given to complete the first questionnaire, participants are given permission to consume the snacks provided on their desks: a bowl of high-calorie and a bowl of low-calorie snacks (not labelled as such or in any other way). After completion of the first questionnaire, the bowls of snacks are collected.

Participants are then asked to complete a second questionnaire. The questionnaire features questions about food consumption and food preferences of the participant and their youngest child as well as the participant’s food consumption during pregnancy. The questionnaire furthermore includes questions about the stressfulness of the stress/control task, chronic stress, participants’ coping behaviours when dealing with stress and about potentially stressful events during the last 3 months as well as during the pregnancy.

At the end of the experimental session, a final saliva sample is collected (approx. 80-90 mins after start of stress / control task).
Before receiving their payment, participants are debriefed: It is explained that the snacks provided differed in calorie content.
Randomization Method
Experimental sessions are pre-assigned to the four experimental conditions ensuring balance regarding day of the week and time of day. Participants indicate their availability for sessions without knowing about the pre-assignment and are assigned to sessions solely based on scheduling concerns.
Randomization Unit
Randomization is done at the experimental session level.
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
15 experimental sessions
Sample size: planned number of observations
approximately 200 participants
Sample size (or number of clusters) by treatment arms
3-4 experimental sessions / approx. 40-60 participants
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
University of Edinburgh School of Economics Ethics Committee
IRB Approval Date
2018-09-25
IRB Approval Number
N/A
IRB Name
Ethics Committee at the European University Institute
IRB Approval Date
2018-09-17
IRB Approval Number
N/A
Analysis Plan

Analysis Plan Documents

Pre-Analysis Plan: Maternal Stress and Food Preferences

MD5: f9ed4fcf79a74f419ceb2f8dcdb707dc

SHA1: af2f309beea4327f7df728214e8035a05f574f8e

Uploaded At: October 13, 2018

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
Yes
Intervention Completion Date
October 19, 2018, 12:00 +00:00
Data Collection Complete
Yes
Data Collection Completion Date
October 19, 2018, 12:00 +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
15 experimental sessions
Was attrition correlated with treatment status?
Final Sample Size: Total Number of Observations
196 eligible participants
Final Sample Size (or Number of Clusters) by Treatment Arms
Stress Task & Simple Shopping Choice: 58 eligible participants Stress Task & Complex Shopping Choice: 49 eligible participants Control Task & Simple Shopping Choice: 49 eligible participants Control Task & Complex Shopping Choice: 40 eligible participants
Data Publication

Data Publication

Is public data available?
No

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Program Files

Program Files
No
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials