Investigating infant expectation on object search tasks.
Infant, behaviours, theory
The current study aims to distinguish between Piaget’s (1954) theory of object understanding, highlighting the
role of object permanence on A not B task performance, and Diamond’s (1985) theory highlighting the role of
motor demands and lack of ability to inhibit habitual behaviours during the task. These two theories differ in
their predictions for the expectations of the infants taking part, with Piaget (1954) predicting that infants’ lack
of object permanence causes poor performance on the task and Diamond (1985) predicting that infants
understand the movement of objects and a lack of inhibition of habitual behaviours cause error in performance.
We tested 15 nine-month-old infants on a looking version of the A not B task. The use of impossible and possible
outcomes was also incorporated on B trials, with the object being revealed from either the correct or incorrect
location (e.g., see Ahmed & Ruffman, 1998). Infant first look direction, accumulated looking time during trials
and the number of social looks initiated post-outcome, were used as measures. We found significant evidence
of the ‘AB’ error during trials, with an significantly increased number of incorrect first looks on B trials. There
was also a descriptive pattern showing surprise at object location reveals with increased number of social looks
during B compared to A trials, though this was not significant. Accumulated looking analysis showed that infants
looked longer on A than B trials, suggesting that infants expected the object to be in location B on B trials,
demonstrating infants’ ability to understand objects and supporting Diamond’s (1985) theory. However,
implications for a small sample size and presence of individual differences on interpretation of looking time data
are discussed. Implications in theory and future research are suggested and overall, results provide support for
the application of Piaget’s (1954) theory and suggest that infants have limited object understanding based on
their displayed expectations during testing.
Leah Murphy
3.1. Participants
In this study, 15 participants took part, aged 8 months and 12 days to 9 months and 27 days old (M = 9
months and 3 days, SD= 11.3 days). Six further infants were excluded from data analysis as they became too
fussy to complete the study. Participants were recruited from the Lancaster Baby lab database, along with the
Lancaster Baby lab Facebook page and were also recruited via word of mouth from guardians taking part in the
study.
3.2. Materials
The video stimuli were created using Canva software (Canva.com, 2023) and was uploaded onto ‘Habit
2’ software (see Oakes et al., 2019) to display the stimuli during testing and to measure the accumulated looking
time of the infant participants. The stimuli involved a novel object obtained from the NOUN database (Horst &
Hout, 2016). A camera was used to record the social looks exchanged between the infant and guardian, as well
as the direction of the infants’ first looks during testing.
3.3. Design
This study had a within-subjects design, with all participants being exposed to the same experimental
conditions and the same stimuli. To counterbalance for location effects, half of the participants witnessed A
trials being hidden in the box on the left, whilst the other half witnessed the object being hidden in the box on
the right during A trials. The presentation of the accurate and inaccurate B trials was further counterbalanced
across participants, as half of the participants viewed the inaccurate B trials first, and the other half viewed the
accurate B trials first.
3.4. Ethical approval
Ethical approval for this study was granted by the departmental ethics committee (DEC) at Lancaster
University. Guardians were recruited via their preferred contact method and were sent the participant
information sheet to read before agreeing to take part in the study. A date and time of testing was arranged at
the Babylab building at Lancaster University, via telephone or email. Upon arrival, guardians were presented
with the consent form to sign and initial all points before being allowed to take part. They were also given the
opportunity to ask any questions about the study and were informed that they could withdraw at any time.
After the study, the guardian received a five-pound contribution to travel costs, along with a free children’s book
for the infant, as a reward for taking part in the study. The guardian also received a debrief sheet to read and to
take home, providing them with all contact information of the lead researcher, if they wished to ask any
questions or to withdraw from the study.
3.5. Procedure
The testing took place in a private room within the Whewell building at Lancaster University. The infant
and guardian were sat in front of a computer screen with the infant sat in a highchair positioned directly in front
of the screen, and the guardian sat in a chair to the side, slightly behind the infant (to allow researchers to see
clearly when the infant initiated a social look). The experimenter sat behind a divider at a computer, out of sight
of the infant and guardian. A social engagement video of the experimenter saying, “Let’s hide the blap, can you
find the blap?” was presented to the infants at the start of the experiment and between each trial, to insert
social communication and guide the attention of the infant to the screen before the stimuli were presented. The
infant then watched a series of video stimuli in which a novel object appeared on the screen and moved into
one of two boxes, both boxes were then covered (the object was hidden), and a there was a delay period of five
seconds (see figure 1). After the delay period, both boxes were revealed, and the location of the toy was visible
to the infant. Any movement of the object was accompanied by a sound to guide the attention of the infant to
the object, but this sound was not present when the object was revealed to avoid any leading factors when
measuring infant expectation. Instead, the occluders made a simple “whoosh” sound when they were removed,
to ensure the infant was paying attention. After five identical A trials, the object was then hidden in the second
location and the process was repeated consisting of six B trials. However, during the B trials, the object was
hidden in the second location, but was either revealed to be in the correct (accurate) or incorrect (inaccurate)
location (see figure 2). This variation in outcome was presented alternately to the infant, with the object being
revealed from the incorrect location for three out of the six B trials. The study lasted for approximately 10
minutes per participant.
Figure 1
Example of A not B task stimuli presentation during A trials or accurate B trials.
Figure 2
Example of A not B task stimuli presentation during inaccurate B trials.
3.6. Behavioural coding
Infant looking time was coded online as trial lengths were infant controlled. Each trial ended when the
infant looked away for four seconds. As this controlled the trial length, this was not double coded as this
inherently will lead to a high agreement level. For the coding of infant first look and number of social looks, the
videos recorded of the participants were saved and uploaded onto Microsoft OneDrive to be offline coded. First
look was defined as the direction that the infant first looked towards once the occluder was removed and the
object was revealed. On trials where the infant was not looking as the occluder was removed, the first look was
defined as the direction in which they looked once their gaze returned to the screen. The first look direction was
coded as correct and incorrect. The number of social looks initiated by the infant per trial was also measured
during coding, defined by the infant turning towards the guardian during each trial after an outcome was
revealed. Twenty percent of the videos were dual coded and there were no discrepancies between researchers
during the dual coding process for first looks (r = 1, p<0.01) or social looking (r= 1, p<0.01).
Lancaster University
2023
Alicja Kowalska
Open
None
Text/Word.doc
English
Text
Murphy2023
LA1 4YW
Cerebral Lateralisation for Emotion Processing of Chimeric Faces in Individuals with Autism Spectrum Disorder
autism spectrum disorder, cerebral lateralisation, emotion processing, adults, children, chimeric faces task
Many studies have suggested that typical lateralisation for emotion processing tasks, such as facial emotion recognition, is lateralised to the right-hemisphere, with different emotions eliciting differing strengths of lateralisation (Bourne, 2010). However, there has been much debate as to the lateralisation of individuals with autism spectrum disorder (ASD) (Ashwin et al., 2005; Shamay-Tsoory et al., 2010). This study assessed the cerebral lateralisation of 30 adults with ASD, five children with ASD, 435 neurotypical adults and ten neurotypical children in a chimeric faces task, and aimed to identify whether the atypical lateralisation seen in children with ASD persists into adulthood (Taylor et al., 2012). Furthermore, the study aimed to identify whether lateralisation strength is affected by the emotion of the facial stimuli. No emotion- or age-related change in lateralisation was found, however, participants with ASD demonstrated a weaker right-hemispheric lateralisation compared to neurotypical participants. Therefore, this study supported the concept that individuals with ASD show atypical lateralisation which persists into adulthood, however, no evidence was found to support the concept that different emotions elicit different strengths of lateralisation.
Alexandra Crossley
Method
Participants
Data from a total of 481 participants with native level English proficiency (or age expected language development in children), normal or corrected-to-normal vision and no history of neurological disease or hearing loss were analysed for the current study (Table 1). Participants in the group ‘adults with ASD’ (N = 30; age: M = 30.17, SD = 9.85) were recruited through adverts on social media, through Prolific Academic (www.prolific.co), and through word of mouth. Participants in the groups ‘children with ASD’ (N = 5; age: M = 6.8, SD = 1.48) and ‘neurotypical children’ (N = 11; age: M = 7.0, SD = 1.90) were recruited through primary schools and word of mouth (Brooks, 2023), and parents of potential child participants were required to email a researcher to express their interest in participation. Participants in the group ‘neurotypical adults’ (N = 435; age: M = 29.44, SD = 8.03) were recruited through Prolific Academic (www.prolific.co) as part of a larger online behavioural laterality battery (Parker et al., 2021). Of the 481 participants who took part in the study, 32 were excluded during the data cleaning process (see Table 1 and Data Analysis for further information).
Measures
As part of the study, a series of questionnaires were administered to collect information about the participants to ensure that individual differences could be accounted for. Participants were asked to complete the study and its associated questionnaires and tasks prior to beginning the main chimeric faces task, and were requested to use a desktop or laptop computer for the entirety of the study. For the ‘neurotypical children’ and ‘children with ASD’ groups, parents were asked to complete the questionnaires on behalf of the children and were asked to be present for the tasks, which were completed during a Microsoft Teams call with a researcher.
The study was completed online using the Gorilla Experiment Builder (www.gorilla.sc), a cloud-based tool for collecting data in the behavioural sciences.
Demographic Questionnaire
The demographic questionnaire asked participants their age, gender, length of time in education (in years), language status, two questions assessing handedness (“Which is your dominant hand? / Which hand do you prefer to use for tasks such as writing, cutting, and catching a ball?”) and footedness (“Which foot do you normally use to step up on a ladder/step?”), and two eye dominance tests (Miles, 1929; Porac & Coren, 1976). Participants were also asked whether they had a diagnosis of any developmental disorders, including ASD, dyslexia, attention deficit hyperactivity disorder or a language disorder (such as 'developmental language disorder' or 'specific language impairment'). For each diagnosis, participants had the option to answer “Yes”, “No”, or “Prefer not to say”, with the exception of ASD which also had the option to answer “No but I am self-diagnosed”. At this point, participants were sorted into their groups based on age (‘children’: five- to 11-years-old; or ‘adults’: 18- to 50-years-old) and ASD diagnosis (‘with ASD’, or ‘neurotypical’). Adults with a self-diagnosis of ASD were included in the ‘adults with ASD’ group.
Edinburgh Handedness Inventory
The Edinburgh Handedness Inventory (EHI; Oldfield, 1971) was administered to provide a scaled score of handedness. Adult participants were asked to score ten daily tasks on a five-point Likert scale based on which hand they preferred to use during each task (“Left hand strongly preferred” = 2, “Left hand preferred” = 1, “No preference” = 0, “Right hand preferred” = 1, or “Right hand strongly preferred” = 2). These tasks included daily activities such as writing, brushing teeth, and opening a box. The EHI was scored by combining the direction and exclusiveness of the hand preference. Two totals were created: one of right-hand preference and one of left-hand preference. The difference was then found by subtracting the left-hand total from the right-hand total. This was then divided by the total score of both hand preference scores and multiplied by 100 (i.e., 100 x (right-hand total – left-hand total) / (right-hand total + left-hand total)). Final EHI scores ranged from -100 to +100, with positive scores indicating right-handedness, and negative scores indicating left-handedness. Child participants were not required to complete the EHI questionnaire.
Lexical Test for Advanced Learners of English
A version of the Lexical Test for Advanced Learners of English (LexTALE; Lemhöfer & Broersma, 2012) was provided to assess the participants’ level of proficiency in English. Within this, adult participants were shown 60 written stimuli comprised of English words and pseudowords (words that follow the orthographical and phonetic rules of the English language and are pronounceable but are otherwise nonsense words, e.g. ‘proom’) and asked to assess whether each word was an existing English word or not. Scores of the test were collected by averaging the percentages of correct answers for English words and pseudowords, with final scores ranging from 0-100. Child participants were not required to complete the LexTALE task.
Autism-Spectrum Quotient (Short Version)
An abridged version of the Autism-Spectrum Quotient (AQ-Short; Hoekstra et al., 2011) was used to provide a measure of ASD traits. Participants with ASD were asked to rate 28 statements on a four-point Likert scale based on their level of agreement, with each answer accruing a different number of points (“Definitely agree” = 1, “Slightly agree” = 2, “Slightly disagree” = 3, or “Definitely disagree” = 4). On items in which “Definitely agree” represented a characteristic of ASD, the scoring was reversed. The scores for each question were totalled, with potential scores ranging between 28 (no ASD traits) to 112 (full inclusion of all ASD traits). Scores above 65 indicated ASD traits to a diagnosable degree. Neurotypical participants were not required to complete the AQ-Short questionnaire.
Procedure Lateralisation for Facial Emotion Processing Task
A chimeric faces task was used to assess lateralisation for facial emotion processing.
Stimuli. The chimeric faces stimuli were created by Dr Michael Burt (Burt & Perrett, 1997) and provided by Parker et al. (2021).
A collection of 16 different facial stimuli were created by merging two photographs of a man’s face depicting one of four emotions (‘happiness’, ‘sadness’, ‘anger’, or ‘disgust’) vertically down the centre of the face and blended at the midline (see Figure 1 for an example). Each emotion was paired either with itself, causing both hemifaces of the facial stimuli to match in emotion (a ‘same face’), or with a differing emotion, causing both hemifaces of the facial stimuli to be different (a ‘chimeric face’). Of the 16 stimuli, 12 were ‘chimeric face’ and four were ‘same face’.
Task. Each trial began with a fixation cross shown for 1000ms, followed by the face stimuli for 400ms. Participants then recorded which emotion they saw most strongly by clicking the corresponding button from a choice of the four emotions (Figure 2). For the children, emoticons were used instead of written words (Oleszkiewicz et al., 2017) (Figure 3). A response triggered the beginning of the next trial, with a time-out duration set at 10400ms after which the next trial was triggered automatically. Response choice and response times were recorded.
The task was split into four blocks of trials with a break between each block. Stimuli were presented in a random order and shown twice in each block, resulting in the participants being shown 32 stimuli per block and a total of 128 within the whole task.
Participants were familiarised with the stimuli at the start of the task, with the ‘same face’ stimuli being shown alongside a label explaining which emotion was being presented, to ensure they could recognise the emotions. A practice block was given at the start of the task to ensure participants knew how to complete the task, using the emotions ‘surprise’ and ‘fear’.
Additional Measures
As data collection also included tasks for other studies, participants were also asked to complete a version of the Empathy Quotient – short (Wakabayashi et al., 2006), and undertake a dichotic listening task and its associated device checks (Parker et al., 2021). As these items were not part of the main study, participants were asked to complete these following the completion of the main study and its associated questionnaires and tasks, to ensure any findings from the study were not due to the additional measures.
Laterality Index
A laterality index (LI) for each participant was calculated using the same method as Parker et al. (2021) by finding the difference between the number of times the participant chose the right-hemiface emotion and the left-hemiface emotion. This was then divided by the total number of times they chose either the right- or left-hemiface emotion, and multiplied by 100 (i.e., 100 x (right hemiface – left hemiface) / (right hemiface + left hemiface)). Scores ranged between -100 and +100, with a negative LI indicating a left-hemiface bias, and thus, a right-hemispheric dominance, and a positive LI showing the opposite.
Data Analysis
Participants who scored less than 80 on the LexTALE task were removed as it was deemed their understanding of the English language was not strong enough and may cause issues with understanding the instructions (Parker et al., 2021). Furthermore, all trials with a response time faster than 200ms were removed as it was suggested that responses at this speed were too quick to have been based on the processing of the stimuli (Parker et al., 2021). In addition to this, outlier response times for each participant were removed using Hoaglin & Iglewicz's (1987) procedure. Within this, outliers were any response times 1.65 times the difference between the first and third quartiles, below the first quartile or above the third (e.g., below Q1 – (1.65 x (Q3-Q1)), and above Q3 + (1.65 x (Q3-Q1))). Following the removal of all outlying trials, any participant with less than 80% of trials remaining were removed. In addition to this, participants who scored less than 75% on ‘same face’ trials (trials in which both hemifaces depicted the same emotion) were noted, because emotion processing is an area of difficulty for individuals with ASD. Within this, three participants in the ‘children with ASD’ group (60%), three participants in the 'neurotypical children’ group (27.27%), four participants in the ‘adults with ASD group (13.33%), and 30 participants in the ‘neurotypical adults’ group (7.41%) scored less than 75% on ‘same face’ trials, suggesting they had difficulties identifying the emotions.
To address the hypotheses, a linear model was performed using LI as the outcome and group (‘ASD’ or ‘neurotypical’), age (‘adult’ or ‘child’) and emotion (‘happy’ and ‘angry’, or ‘sad’ and ‘disgust’) as the predictors, including interactions between each predictor (Group x Age; Group x Emotion; Age x Emotion; and a three-way interaction, Group x Age x Emotion).
Lancaster University
5th September 2023
Open
None
.csv
English
Data
Crossley2023
LA1 4YF
What person attributes influence the comprehension of written health information? A scoping review and critical appraisal
health literacy, comprehension, person attributes, health outcomes.
Increasingly, individuals are required to be actively involved in their healthcare. To do so successfully, individuals need to possess the skills and resources to be able to access, understand, and apply health information. Health communication guidance proposes that health information is not understood due to the mismatch between adults average literacy skills and the literacy skills required to comprehend health information. To tackle this, the use of plain language, such as shortening sentences and removing jargon, is promoted. Policies, however, do not commonly consider the impact of person attributes, such as age, education, and gender, on the comprehension of health information. To understand the nature and scope of current research, and whether person attributes do have an impact, a scoping review was conducted. The search strategy yielded 5,459 articles which were then screened, resulting in a final sample of 99 studies. Quantitative analyses and a critical appraisal revealed three main findings: (1) the research is heterogenous and evolving; (2) person attributes are not commonly used in analyses; and (3) when person attributes are included, the effects on comprehension vary. The findings and implications of this review have the potential to influence how future research is conducted, and crucially inform policies about the importance of person attributes on the comprehension of health information.
Charlotte Betts
Stage 1: Identify the Research Question
The current research is an updated scoping review, building upon earlier work by Davies et al. (in preparation), which seeks to answer: What person attributes affect or can be predicted to affect the response of individuals to written health information?
Table 2
Form Developer: Rebecca A. James
Search strategy methods
Strategy
Method
Bibliographic
Searched the following: Cumulative Index to Nursing and Allied Health Literature (CINAHL); PsycINFO; PubMed; and Web of Science (WoS).
Journal
Obtained all sources from the following journals between 2018-11-08 and 2023-05-05: Patient Education and Counselling; Health Communication; and Journal of Health Communication.
Author
Obtained sources from the following authors between the dates 2018-11-08 and 2023-05-05: TC Davis; Dan Morrow; Chiung-Ju Liu; Michael Paasche-Orlow; Lisa Soederberg Miller; Rima Rudd; and Michael Wolf.
Reference
Once the full text of the bibliographic, journal, and author searches were complete, the reference lists of the included items were examined to locate new and possibly relevant articles.
Stage 2: Identifying Relevant Studies
Sources were identified using four methods: (1) bibliographic search; (2) journal search; (3) author search; and (4) reference search. Grey literature was not searched due to concerns with the quality of the literature and possible time constraints. Excluding the reference search, all articles were published between 8th November 2018 to 5th May 2023. Details of each of the search methods are outlined in Table 2.
Stage 3: Study Selection
Once articles were imported to Rayyan, a free online software application for conducting reviews (Ouzzani et al., 2016), duplicate articles were identified and removed. Then articles went through a title and abstract screening whereby articles which did not include the following were excluded: (1) a measure of understanding, comprehension, or readability; (2) a quantitative outcome; (3) populations who are typically developed; (4) presentation of health information; (5) present original data (excluding reviews); and (6) presented in English or English was a first language.
The exclusion criteria (Table 3) enabled the final sample of studies to be focussed and relevant to the review. Included articles were then read in full and the same exclusion criteria was applied. Articles which passed the full-text screening were then examined to identify relevant studies from the reference lists, and these references then underwent the same screening process outlined above. Following best practice recommendations (Levac et al., 2010), study selection was conducted by myself and TM (a MSc student) to reduce the chance of bias. Further, regular training and meetings took place (with TM and supervisor RD) to become familiar with the process and to discuss and resolve conflicting decisions between researchers.
Table 3
Exclusion criteria for study selection
Exclusion Criteria
Reasoning
Not a measure(-s) of understanding, comprehension, or readability, metacomprehension, or recall
Articles which do not measure understanding, either directly, or indirectly, and do not measure readability of texts, are not relevant to the current review.
Not quantitative outcomes
Quantitative data is needed to understand average associations between the variation of person attributes and comprehension responses.
Not typical development (excluding participants presenting cognitive or language impairments)
Need to first understand how responses to health information varies within a typical population. Future research should be more inclusive to see how response varies in the whole population.
No presentation of health information.
The present review is concerned with comprehension and response to written health information.
Not original data (rather than reviews).
Although reviews themselves are not targets for review, they will be identified as potentially informative.
Not English or second language speakers of English.
There is limited information regarding how comprehension responses to text may be different or similar in different language, further, text properties may differ.
Stage 4: Data Charting
Articles were classified as being either an experimental, readability, or review article and as this paper focusses on research investigating the effects of person attributes, only experimental articles are analysed and reported. TM analysed and reported readability studies. Data extraction was completed so that information about the nature and characteristics of the study could be recorded. Data extraction was achieved by entering information (Table 4) into an online Qualtrics form which was developed and used by Davies et al. (in preparation) in their scoping review, which allowed for systematic extraction of information regarding the characteristics, methods, and findings of each study. To ensure that data extraction was reliable, a sample of studies were charted in parallel by myself and TM and were checked by RD for consistency.
Table 4
Characteristics that will be extracted from experimental studies for data charting.
Form Developer: Rebecca A. James
the article title
the article DOI, if available
the article authors
the article year of publication
the location of data collection (location may be inferred by author affiliation, or reported in article text concerning the regional or national source of health texts, or the locality of participant recruitment)
information about the composition of the participant sample (e.g., healthy adults, patients, etc.)
the number of participants
individual differences measures, if reported (e.g., gender, age, etc.)
text type (the type of the health information text sampled, e.g., website, medicine information, etc.)
text topic (the topic of the health texts sampled)
text sample size (the number of texts sampled)
if the study involved the manipulation of text properties, information on what linguistic or other features were manipulated, or what intervention was implemented (e.g., variation in organization or structure, in the inclusion of pictures, in readability, format, or other)
what test of comprehension was conducted (e.g., verbal or written summary, true/false question, open-ended questions, multiple-choice questions, cloze, recall, etc.)
what outcome measure was analysed (accuracy, or other)
Stage 5: Collating, summarising and reporting the results
Data charting resulted in the creation of a database of detailed information about the nature and scope of each article. To effectively make sense of such information, the original database of information was organised using thematic labels (Table 5). For example, the thematic label leaflet would be applied to articles which referenced handouts of medical information as pamphlets, leaflets, and brochures. This process enabled greater ease and clarity to conduct quantitative analyses and to provide a textual commentary of the findings. Quantitative analyses include frequencies and distributions of study characteristics observed in the sample, in addition to evidencing what direction of effect person attributes had on responses to health information. Directionality of the results, as opposed to reporting significance is deemed appropriate as the reporting of significance in reviews is misleading (McKenzie & Brennan, 2019). Following the synthesis and quantitative analyses, a critical appraisal of the evidence was conducted. Although this stage is optional for scoping reviews (Tricco et al., 2018; Levac et al., 2010), it was considered necessary to provide a sense-making of the conclusions we can reach given the synthesis of evidence. The appraisal followed guidance from the Synthesis Without Meta-Narrative (SWiM) guidance (Campbell et al., 2020) and Realist And Meta-narrative Evidence Syntheses: Evolving Standards (RAMESES) publication standards (Wong et al., 2013). Such guidance provides a framework
Form Developer: Rebecca A. James
for the critical appraisal to comprehensively answer the research question, and discuss the traditions, trends, and value of research. Unlike other reviews such as systematic reviews, formal assessment tools such as the Cochrane Risk of Bias tools will not be used as this review does not focus on examining randomised control trials and the research is too heterogenous to appropriately apply such tools (Levac et al., 2010).
Table 5
Thematic labels for experimental studies
location
text type (e.g., consent form, decision aid)
topic or health area (e.g., arthritis, cancer)
intervention (e.g., counselling, drug)
[study] design (e.g., illustration type, text readability)
[study] implementation (e.g., different data visualizations, different organisation)
outcome (e.g., comprehension, knowledge)
[outcome] measure (e.g., multiple choice question, self-rated)
individual differences (e.g., age, gender)
Lancaster University
11/09/2023
Oliver Powell
Unsure. Contact Dr. Rob Davies.
In part, in collaboration with TM. Supervised by Dr. Rob Davies
Data.csv and Text.doc
English
Scoping Review
Betts2023
LA1 4YW
How to add records to the system
Prospect theory, gain/loss framing, intermediate audience, communication research, health communication, vaccination
Prospect theory predicts how people react to gain or loss-framed outcomes in dilemma situations, where the potential consequence of the choice is framed as a gain (e.g., lives saved) or as a loss (lives lost). This gain-loss framing communication strategy, derived from the theory, has been applied in many contexts, from promoting the use of reusable coffee mugs to vaccination compliance, with loss-framed appeals being found generally to be more persuasive than gain-framed appeals in the context of promoting vaccination. The current study focused on exploring whether these well-established effects persist when an intermediate audience is exposed to gain/loss-framed messaging, using influenza (flu) vaccination intentionality as an outcome. Intermediate audiences refer to those who are evaluating the gains and losses from the message on behalf of someone else (the ultimate audience), while normal audiences are those making decisions on their own behalf. Two hundred participants were recruited for an online, between-subject study, in which participants were split into two audience conditions and within which they were further split to view a gain-framed or a loss-framed message. Their subsequent behavioural intentions were measured as the outcome, with age as a potential moderating factor (and emotional attachment as a potential mediator exclusively for the intermediate audience condition). Results indicate that neither age nor emotional attachment are significant moderators or mediators. Loss-framed appeal enjoyed a persuasive advantage over the gain-framed appeal only in the intermediate audience condition. Possible interpretations of results, along with potential further directions of research, are discussed.
Wai Man Ko
To test the outlined hypotheses, our current study took the form of an online Qualtrics questionnaire (see appendix B for questions) where the questionnaire would introduce participants to one of the audience conditions and view the appropriate version of the manipulated message before moving on to answering some items measuring their behavioural intention and emotional attachment. The study has a 2 (intermediate/normal audience condition) X 2 (gain/loss-framed appeal) design with emotional attachment as a potential mediating variable for the intermediate audience condition and behavioural intention as the outcome variable for all audience conditions.
Participants
We recruited 200 healthy adults based in the UK on Prolific, an online research participant recruitment platform. Participants have provided consent and completed the study remotely with their personal devices. Their unique Prolific ID was used in this study as the only identifier, which cannot be traced back to them personally. Participants were compensated monetarily for their participation.
We randomly assigned our participants to one of the four audience conditions with 50 participants each: the normal gain-framed condition, the normal loss-framed condition, the intermediate gain-framed condition, and the intermediate loss-framed condition.
Questionnaire design
Consent
The participant gave consent to participate in the study with the Qualtrics consent element so that participants can check a box for each item. There were seven items that the participants had to check one by one before commencing the study. Responses which failed to provide a full response in the consent item would be removed from the study.
Demographics
For demographics, we have recorded the participants' age and gender for the records. As mentioned, age was also analysed as a moderator as part of our analysis. We have also recorded their Prolific IDs to ensure completion and arrange payment.
Settings of the study
After giving demographic information, participants were introduced to a small piece of information that gave them the context of this study. In normal audience conditions, participants were told that someone had sent them an ad about the flu vaccination, which refers to the manipulated message they will soon view. While for the intermediate audience, on top of the information that is revealed to the normal audience, they were exclusively told that they were a manager in a small town's paper company, which gives them the role of an intermediate audience (manager) who must evaluate the later presented message on behalf of other parties (employees) with themselves irrelevant to the gains and losses.
Material
We have chosen flu vaccination as our topic malady for the manipulation messages as COVID vaccines, as used in recent studies, are perhaps less relevant in what is generally thought of as the post-COVID era. Flu vaccinations, unlike many other vaccines, remain relevant to the major population and most age groups. To allow a closer resemblance to real-world settings and increase the generalisability of the results, we have made unofficial Facebook posts that claim to be from the NHS as the message format. Participants were informed that the graphics were not an actual Facebook post from the NHS but rather a material used solely for this study. See Figure 2 for an example, and appendix A for the complete set of stimuli presented to the participants in the study.
Audience condition. Figure 2 is the gain-framed version of the message from the normal audience condition. In normal audience conditions, the message communicates directly to the participants, stating the potential pros or cons for the participants when the participants decide to vaccinate or not vaccinate. In this condition, it is assumed that the participants evaluated the message on their behalf and nobody else's. While on the contrary, the intermediate audience condition communicates a slightly different message. The "you" in the message is replaced by "your employees". The purpose of this is to highlight that the participants evaluate this message as an intermediate audience (the manager), deciding whether they would recommend the vaccine to somebody else (the ‘ultimate audience’) given the outlined potential gains and losses, while the gains and losses remain irrelevant to the participants personally.
Message framing. The figure is a gain-framed message, and as mentioned, it follows the logical flow of "if you vaccinate, good things will happen". As we can see in Figure 2, if the recipient vaccinates, then according to the text, he/she would have a reduced chance of infection and a reduction in the duration and severity of the symptoms. The lost-framed version of the message follows the logical flow of "if you do not vaccinate, bad things will happen." So, in contrast to figure 2, the lost framed messages would say if the recipient does not vaccinate, he/she would have an increased chance of infection and increase in duration and severity of the symptoms. The two messages communicate the same reality and are logically equivalent. Hence, any differences between the groups can be attributed to the message framing.
Check questions.
After viewing the message, the participants were asked two questions regarding the ads content before moving on to later questions. The check questions were designed to be simple reading comprehension questions that check whether the participants attended to the message in the reading process. We have removed all responses failing to provide a correct answer in either one of the questions.
Behavioural intention
After viewing the framed messages, we have several Likert scale 7-point agree-disagree items used to measure the behavioural intention of the participants. However, given the audience condition differences and hence the potential differences in the decision-making process, behavioural intention for the two types of audience is defined differently. For the intermediate audience condition, behavioural intention is defined as "the intention to recommend/promote behaviour to the ultimate audience (employees)". While for the normal audience conditions, we measure their intention to get the vaccination for themselves. Both audience conditions responded to six items probing their behavioural intentions. In the normal audience condition, participants were asked how likely they would be to get the flu jab, how urgent they thought it is, and whether they would likely plan to get a flu jab after viewing the message. There are also items with reversed wordings asking whether they think getting a flu jab is NOT urgent. The intermediate audience was asked how likely they are to recommend the flu vaccine to their employees and how urgent and necessary they believe the vaccine is to their employees. (See the appendix for the complete set of questions.)
Emotional attachment
As mentioned, there are speculations revolving around the involvement of relational dynamics and relevant emotions in the intermediate audience. Therefore, we have arranged a set of questions probing the participant's emotional attachment towards the employee exclusively for the intermediate audience condition. There were four questions in total in this part of the study, which focused solely on the participants' sense of protection towards the employee, asking to what extent the participants thought that the vaccine was necessary for the employee's own good and well-being, and to what extent were the participants eager to protect them; an item with reversed wordings were also included. (See the appendix for the complete set of questions.)
Method of analysis
We analysed the data using the clm() and clmm() functions from the ordinal package in RStudio using R version 4.1.1. We first confirmed the main effects of message framing and audience conditions using clm(), and then we moved on to analyse the magnitude of random interacting effects of age, question type and individual differences. The reason for choosing cumulative link models (clm) was that the models were designed explicitly for ordinal variables like Likert scales, which predict the probability of each response level, unlike some metric models and prevent type 1 and type 2 errors resulting from forcing ordinal variables onto metric models (Liddell & Kruschke, 2018). As for emotional attachment, given each item was probing quite a different emotion (e.g., sense of responsibility/ sense of protection), we have decided to fit a multivariate ordinal variable using the mvord() function to see if there is a significant difference in the multiple emotional outcomes under different audience condition, after which we investigated if any emotional attachment item was a significant predictor of behavioural intention using another clm model. We have also fitted clm() models including the interaction term between age and conditions predicting behavioural intention to see if age moderates the relationship between message framing and behavioural intention as proposed. Lastly, we have fitted a cumulative link mixed model (clm) to consider the role of potential sources of random effects such as participant differences and question differences in the analyses.
Lancaster University
2023
Hannah Clough
Open
Data.csv
English
Data
Ko2023
LA1 4YF
Does Noise Affect How Children Learn Grammar in the Classroom?
Grammar, Noise, Working Memory
In a classroom environment noise can be a significant impediment, obstructing and distorting essential information being taught. Extensive prior research consistently indicates that noise has a detrimental impact on learning, those who learn in noise retain and comprehend far less information than their counterparts who learn in quiet. To date there are no studies that investigate the effect of noise on learning grammar specifically -the primary aim of the current study is the address this research gap. This paper details our recruitment of 16 children aged 7– 12 through the Babylab database at Lancaster university. This study employed a between participants design, where children completed a three-part audio evaluation, engaged in an artificial grammar paradigm, and a undertook a working memory task. The artificial grammar paradigm was employed as our primary assessment tool, participants were exposed to the grammar either in noise or in quiet. Results were analysed using a multiple regression with total grammar score as the dependent variable and age, gender, condition, and working memory as the independent variables. In contrast the prior research, our results revealed that the effect of the independent variables on the dependent variable was statistically nonsignificant, proving our null hypotheses to be true. These findings suggest that background noise does not affect how children learn grammar in the classroom challenging the existing understanding that noise negatively impacts learning.
Analysis
In order to answer our research questions we will carry out a multiple linear regression using IBM SPSS Statistics (version 28). We will be employing a between participants design where we will examine the effect of background noise (noisy and quiet) on total grammar score. Our additional independent variables will be working memory, gender and age. If we find a statistically significant result with regard to grammar score then we will be conducting a post hoc test on grammar score breaking them down into aX and Yb in order to determine the difference between the two types of grammar.
Ashlynn Mayo
Participants
16 children aged 7-12 years old participated in this study, unfortunately due to technical issues 5 participants’ data were excluded leaving 11 children’s data to be included in the analysis (M=8.64, SD=1.63, female=7, male=4). Children were recruited through the Lancaster University Babylab database and by flyers posted on social media and local community. A requirement of the current study was that children be English speaking monolinguals, this is because an abundance of research has indicated that those who can speak two or more languages are at a far greater advantage when it comes to new language acquisition (Antoniou et al., 2015). Therefore, in order to control the likelihood of extraneous variables such as this we ensured all participants were English speaking monolinguals only.
Furthermore, children were also required to have normal vision or corrected to normal vision. To rule out hearing loss all children had to pass an otoscope inspection, a tympanometry test, and a pure tone hearing screening at 20dB in the standard frequencies (250Hz-8kHZ).
The current study employed a between participant design whereby subjects were allocated to a condition based on their age and gender -age was categorised into 7-9 and 10-12 in order to ensure that there were as equal an amount of males and females in each condition over all ages. It is crucial for the validity of the study that children are only exposed to the artificial grammar paradigm once or data will be rendered unreliable as they will have an unfair advantage over the other participants.
Ethics for the current study have been obtained from the Departmental Ethics Committee (DEC), Psychology Department at Lancaster University.
Materials
This study was conducted within a double walled soundproof chamber at Lancaster University’s PELiCAN lab where the participant sat at a desk with a monitor placed in front of them. A secondary researcher was present in the lab for health and safety purposes.
Consent and assent forms, a background questionnaire on the child’s hearing, audio evaluation results, and task data were all recorded on REDCap (Harris et al., 2009; Harris et al., 2019): a GDPR compliant application for data capture.
Travel compensation was provided: £5 within 40 minutes and £10 for over 40 minutes.
Furthermore, children received a certificate and book of their choosing from the PELiCAN lab.
The audio evaluation
This study was comprised of three sections: an audio evaluation whereby an otoscope examination, tympanometry test, and audiogram using Affinity Suite were conducted. During the audiogram participants wore headphones and had a handheld button that they pressed when they heard the pure tone sounds.
The Artificial Grammar Paradigm
After passing the hearing evaluation the children completed an artificial grammar paradigm previously used by Torkildsen et al. (2013) consisting of two grammatical forms: aX and Yb. The paradigm was presented in the form of an alien game whereby the children helped an alien learn a new language. We presented the paradigm in this format in order to increase engagement; children are motivated by the colourful and curious nature of a game (Blumberg
et al., 2019) and therefore we are far more likely to obtain more data (less drop outs due to fatigue and boredom). This task was created in PsychoPy and hosted by Pavlovia.
The background noise
In order to imitate the background noise of a classroom speech shaped noise (SSN) (e.g. Leibold et al., 2013) was emitted through a speaker on the back wall of the booth behind the child. The background noise speaker was 180 degrees on the azimuth, and the target speaker was 0 degrees on the azimuth. Background stimuli was calibrated so that for the quiet condition the stimulus was emitted at 35dB and for the noisy condition it was played at 65dB.
The n-back Test of Working Memory
Lastly, we conducted the 1-back test of working memory (Owen et al., 2005) which was also created on PsychoPy and hosted by Pavlovia
Procedure
Prior to the commencement of the study guardians gave informed consent (See Appendix C), if the child was 11 or older they gave informed assent in addition to this (See Appendix D). Guardians were then asked to complete a short background questionnaire pertaining to their child’s hearing (See Appendix H). Whilst they completed these forms the researcher began the study inside the booth; using Affinity suite it was ensured that the microphone inside the booth was turned on in order for the guardian to be able to hear what was going on inside the booth by using the headphones places outside the booth. As aforementioned, the audio evaluation consisted of three tests, these were administered in the booth by the researcher and took up to 15 minutes. Firstly, an ear inspection was conducted using an otoscope, participants were required to have clear ears free of perforations and/or any infection. Secondly, a tympanometry test was conducted whereby participants must have passed with type A (normal) results. Lastly a pure tone hearing screening was conducted at 20dB in the standard frequencies (250Hz-8kHZ). The researcher left the booth for the audiogram in order to run the program on the desktop outside the booth while the child remained inside the booth.
The task consisted of 11 blocks comprised of 4 exposure items and 2 test items, before the test portion children were exposed to 4 examples of what is expected of them, they had to get these right in order for the software to move onto the test phase. If children did not get these right the researcher explained and promoted them to pick the correct answer. Children were required to press ‘x’ on the keyboard for right and ‘n’ on the keyboard for wrong, answers were saved and recorded automatically on Pavlovia. The software was run by the researcher from outside the booth and was mirrored onto the desktop inside the booth.
Lastly, we conducted the 1-back test of working memory (Owen et al., 2005), where children were exposed to a number of animal sounds and were required to record weather the stimuli was a new sound or one they had heard before, ‘x’ represented repeated sound and ‘n’ represented a new sound, participants had to ensure they made a button press after each noise. Once all tasks were completed the researcher collected the child from inside the booth and a short verbal and written debrief was given to the child and guardian. Guardians were given and signed for their travel compensation, and children received a certificate from the PELiCAN lab and were able to choose a book of their liking. Participants were walked back to their car or bus to bring a close to the visit.
Lancaster University
2023
Tejasvita Rajawat
Audred Visaya
Open
None
Text/Word.doc
Data/Excel.csv
English
Data
Mayo2023
LA1 4YF
Hemispheric Lateralisation of Facial Emotion Processing: A Possible Explanation of Atypical Empathetic Responses in Children with Autism Spectrum Disorder
Hemispheric Lateralisation, Emotion Processing, Autism Spectrum Disorder, Empathy, Chimeric Face Task
Existing research suggests that children with autism are endowed with a significant delay in the lateralisation of facial emotion processing (Taylor et al., 2012), and that this delay is associated with some of the social and emotional based deficits that manifest within the disorder. The present study therefore aimed to ascertain the reliability of Taylor et al.’s (2012) findings by determining whether the strength of lateralisation for facial emotion processing differs between children with and without autism, while also determining whether this difference can explain atypical empathetic responses in children with autism. To explore these aims, an online version of the chimeric face task was administered to 11 neurotypical children and 5 children with a diagnosis of autism. The Child Empathy Quotient was completed by parents of all children, and The Autism Quotient – Children’s Version was completed by parents of children with autism. Results indicated that there was no significant difference in the strength of hemispheric lateralisation for facial emotion processing between children with and without autism, and that the strength of lateralisation did not predict a child’s level of empathy, nor did a child’s autism severity. Instead, levels of empathy were best predicted by an individual’s diagnostic status and age. The present study was therefore unable to support the finding of Taylor et al. (2012) or explain empathy deficits in the autistic population. However, the limitations identified in this study help to inform future research on the relationship between the lateralisation of facial emotion processing and empathy.
Lydia Brooks
Participants
Participants were recruited from mainstream primary schools, wrap around care settings and specialist educational provisions in the Lancashire area, as well as via social media. A total of 22 parents completed the required questionnaires on behalf of their children, out of which 17 parents arranged a date and time for their child to complete the chimeric face task. One child, who is non-verbal and has received a diagnosis of ASD had difficulty completing the task and selected responses impulsively without looking at, or taking sufficient time to consider, the facial stimuli and the emotion it depicted. For this reason, the chimeric face task was terminated prior to completion and the child’s data was not included in the analysis.
The final sample of participants consisted of 16 children aged between 5- and 10-years-old, of which 5 had received a formal diagnosis of ASD (5 boys; Mage = 6.8, SDage = 1.48). One child with ASD had a comorbid diagnosis of hypermobility and sensory processing disorder. All children with ASD were reported to speak English at home, one child was left-hand dominant, and four children were right-hand dominant.
The remaining participants were 11 typically developing children (6 girls, 5 boys; Mage = 7.0, SDage = 1.90), who had not been diagnosed with any neurodevelopmental disorders. One of these children was reported to speak Russian at home, however, is fluent in English. All children in the typically developing group were right-hand dominant.
Design
A two-factor between-subjects experimental design was employed to determine whether the strength of hemispheric lateralisation for facial emotion processing differs between children with and without a diagnosis of ASD. The independent variable for this research question was diagnostic status, a between-subject factor, with two groups; ASD and typically developing. Participants were assigned to one of these groups based on their diagnostic status, which was ascertained by their parent’s responses on the demographic questionnaire. The dependent variable for this research question was the strength of hemispheric lateralisation for facial emotion processing which was measured using the chimeric face task.
A three-factor mixed-subjects predictive correlational design was employed to determine whether a child’s diagnostic status, and strength of hemispheric lateralisation for facial emotion processing can predict a child’s level of empathy. The predictor variables for this research question were diagnostic status, a between-subject factor (typically developing or ASD), and the strength of hemispheric lateralisation for facial emotion processing, a within-subject factor. The outcome variable for this design was empathy, a within-subject factor, measured by the Child Empathy Quotient.
Measures
Demographic Questionnaire
Materials. The online demographic questionnaire (see Appendix A) was comprised of eight questions. Three of which required parents of the participants to input a response, these questions were used to determine the child’s age (in years), month of birth, and year of birth. The remaining questions were multiple choice, and therefore, required parents to select an answer out of 2-4 possible answer options. These questions acquired information including the child’s gender (male or female), dominant hand (left, right or don’t know/no preference), the language used in their home environment (English or other) and diagnostic status (formal diagnosis of ASD or no formal diagnosis of ASD). If the child did not speak English at home, then parents were required to input the language predominantly spoken. Parents who confirmed that their child had received a formal diagnosis of ASD were asked to input any comorbid diagnoses their child had received, so that they could be considered in the analysis. Parents who confirmed that their child had not received a diagnosis of ASD were asked if their child had received a diagnosis of any other neurodevelopmental disorders, this question was used for exclusionary purposes.
Procedure. Completion of the questionnaire took approximately 2 minutes. Following completion of the questionnaire participants were excluded from the study and unable to proceed to the next stage if they did not meet the age criterion, or if they had not received a diagnosis of ASD but had received a diagnosis of another developmental disorder.
The Child Empathy Quotient (Auyeung et al., 2009)
Materials. The Child Empathy Quotient (EQ-Child) is a parent report questionnaire composed of 27 items (see Appendix B) used to measure a child’s level of empathy. This questionnaire was developed by Auyeung et al. (2009) using the adapted version of The Adult Empathy Quotient (Baron-Cohen & Wheelwright, 2004), individual items have therefore been modified and made applicable and relevant to children. The items therefore refer to behaviours, responses or difficulties commonly exhibited or experienced by children, e.g., ‘My child shows concern when others are upset’. Parents had to indicate the extent to which they agreed with each item by selecting one of the following options on a four-point Likert scale; ‘definitely agree’, ‘slightly agree’, ‘slightly disagree’, and ‘definitely disagree’.
The EQ-Child has previously been completed by parents of neurotypical children, and children with ASD, aged between 4- and 11-years-old. The pilot study conducted by Auyeung et al. (2009) yielded findings indicative of a high-internal consistency and good test-retest reliability, and the patterns of results were consistent with those found in adult research (Baron-Cohen & Wheelwright, 2004).
Procedure. All parents were required to complete the EQ-Child, which took approximately 5 minutes. The order the questionnaire items were presented in remained consistent between parents and parents were unable to proceed to the next part of the study before they had provided a response for all 27 items.
Scoring. Parental responses on individual questionnaire items were converted into numerical points and summed together to calculate an empathy score for each child. For the following numbered items; 1, 4, 8, 10, 13, 14, 15, 16, 19, 21, 22, 23, 24, and 25, a response of ‘definitely agree’ equalled 2, ‘slightly agree’ equalled 1, and, ‘slightly disagree’ or ‘definitely disagree’ equalled 0. The remaining items were reverse coded. The maximal attainable empathy score was 54, the higher the score, the more empathetic a child is perceived to be by the adult completing the questionnaire. The scoring method applied in this study is consistent with the scoring method used, and detailed, in Auyeung et al. (2009). See Appendix B for the 27 items, and their corresponding item number.
The Autism Spectrum Quotient – Children’s Version (Auyeung et al., 2008)
Materials. The Autism Spectrum Quotient – Children’s Version (AQ-Child) developed by Auyeung et al. (2008) is a parent report questionnaire composed of 50 items (see Appendix C), used to quantitatively measure autistic traits in children aged between 4- and 11-years-old. The items in the AQ-Child are derived from the Autism Spectrum Quotient – Adult’s Version (Baron-Cohen et al., 2001) and the Autism Spectrum Quotient – Adolescent’s Version (Baron-Cohen et al., 2006), however, they have been revised and adapted to be pertinent to children. The items therefore refer to scenarios and behaviours that children are likely to have experienced or exhibited, e.g., ‘S/he would rather go to a library than a birthday party’. Parents indicated how strongly they agreed with each descriptive statement by selecting one of the following responses on a four-point Likert scale; ‘definitely agree’, ‘slightly agree’, ‘slightly disagree’ and ‘definitely disagree’.
Previous studies have administered the AQ-Child to parents of children with ASD, aged between 5- and 11-years-old (Auyeung et al., 2008). Administration of the AQ-Child has been reported to have excellent test-retest reliability and a high alpha and reliability coefficient.
Procedure:
This questionnaire was only completed by the parents of children with a diagnosis of ASD, all of whom were unable to proceed to the next stage of the study until they had provided an answer for all 50 items. This questionnaire took approximately 5-10 minutes to complete. The order of items remained constant between parents.
Scoring. For each child reported to have a diagnosis of ASD, a total AQ score was calculated. Total scores were calculated by converting responses on the four-point Likert scale into numerical scores and summing them together. For the following items 1, 3, 8, 10, 11, 14, 15, 17, 24, 25, 27, 28, 29, 30, 31, 32, 34, 36, 37, 38, 40, 44, 47, 48, 49 and 50, a response of ‘definitely agree’ equalled 0, ‘slightly agree’ equalled 1, ‘slightly disagree’ equalled 2, and ‘definitely disagree’ equalled 3. The remaining items were reverse scored. The higher the overall score, the greater number of autistic traits exhibited and endowed by the child. See Appendix C for the 50 items, and their corresponding item number.
The Chimeric Face Task
Materials. The chimeric face task is a widely used measure of the lateralisation of facial emotion processing. Chimeric faces are composite visual stimuli that are made by splitting two symmetrically averaged images of a face vertically down the middle and combining them together to depict a different emotional expression in each hemiface. The chimeric faces and the symmetrically averaged images used in this study derive from the work of Michael Burt (Burt & Perrett, 1997; Innes et al., 2016), and are supplied by Parker et al. (2021) via Gorilla Open Materials: https://gorilla.sc/openmaterials/104636.
In the practice trail, two symmetrically averaged images of male faces and two chimeric faces were used, these faces depicted the emotions fear and surprise. A further 12 chimeric faces were used in the experimental trial, which depicted all possible combinations of the emotion’s happiness, sadness, anger and disgust. Four symmetrically averaged images of male faces depicting these emotions were also used. See Figure 2 for the stimuli used in the experimental trial.
Figure 2. The Facial Stimuli used in the Experimental Trial of the Chimeric Face Task
Note. The 16 facial stimuli presented to children during the experimental trial of the chimeric face task, including the four symmetrically averaged faces depicting the emotions happiness, sadness, anger and disgust, and the 14 possible combinations of these four symmetrically averaged faces. Adapted from “A leftward bias however you look at it: Revisiting the emotional chimeric face task as a tool for measuring emotional lateralisation” by B. R. Innes, D. M. Burt, Y. K. Birch, and M. Hausmann, 2016, Laterality: Asymmetries of Body, Brain and Cognition, 21(4-6), p. 649, supplied by “Assessing the reliability of an online behavioural laterality battery: A pre-registered study” by A. J. Parker, Z. V. Woodhead, P. A. Thompson, and D. V. Bishop, 2021, Laterality, 26.
The participants used emotional emoticons to indicate the emotion they believed to be depicted by the facial stimuli. In the practice trial, two emoticons were used, which illustrated the emotions fear and surprise. In the experimental trial, a further four emoticons were used, which illustrated the emotions happiness, sadness, disgust and anger. The emoticons used were taken from Oleszkiewicz et al. (2017), as it was found that children aged between 4- and 8-years-old were able to accurately assign emotions to these emoticons. See Figure 3 for the emoticons used in the experimental trial.
Figure 3.The Emoticon Stimuli used in the Experimental Trial of the Chimeric Face Task
Note. The emoticon stimuli selected by the child participants to indicate which emotion they believed the facial stimuli to be depicting. The emotions depicted by the emoticons, from left to right, are; anger, disgust, happiness and sadness. Adapted from “Children can accurately recognize facial emotions from emoticons” by in A. Oleszkiewicz, T. Frackowiak, A. Sorokowska, and P. Sorokowski, 2017, Computers in Human Behavior, 76, p. 373.
Procedure. The procedure used derives from Parker et al. (2021), however, it has been adapted accordingly for its use with children. The chimeric face task was administered remotely via Microsoft Teams, a video collaboration platform. The virtual meeting was only accessible by the participant and the researcher, via a unique uniform resource locator, meeting ID and passcode. The chimeric face task could be completed on a laptop, computer or electronic tablet, and participants were required to share their screen to allow for the delivery of verbal instructions. Children were accompanied by an adult family member who was asked to refrain from engaging in any verbal and non-verbal communication with their child during completion of the task.
Prior to administration of instructions, participants completed an estimation of screen size. This involved placing a 8.56cm X 5.39cm card onto the screen and dragging a bar until the size of the card on the screen corresponded with the physical card possessed by the participant. This was to ensure that all instructions and stimuli were presented as the same size to all participants. Instructions were administered visually to the child participants, using visual-graphic symbols, example screens and visual stimuli taken from the study, to ensure that the child’s understanding of the task was not compounded by their language ability. The visual instructions were accompanied by verbal instructions, that omitted the use of vocabulary that would not typically be understood by children aged between 5- and 10-years-old. Following administration of the instructions, participants were familiarised to two symmetrically averaged faces depicting the emotions fear and surprise, and their corresponding emoticons, which would be used in the practice trial. During completion of the practice trial participants were exposed to each symmetrically averaged face and each chimeric face, twice, meaning they were exposed to a total of 8 stimuli. The practice trial was employed to acquaint the child to the procedure used in the experimental trial.
Following the practice trial, participants were familiarised to the symmetrically averaged faces that comprised the chimeric faces used in the experimental trial, as well as their corresponding emoticons. These faces, and emoticons, depicted the emotions happiness, sadness, anger and disgust. Participants were verbally informed of the emotion depicted by stimuli and were instructed to click ‘next’ or indicate to their parent when they felt they had familiarised themselves with the stimuli presented. Participants were familiarised to the stimuli to ensure they knew which emotion each face and emoticon represented. The experimental trial was composed of four blocks. In each block the participants were exposed to the four symmetrically averaged faces, and the 14 chimeric faces, twice, meaning they were exposed to 32 stimuli per block, and 128 stimuli in total. Participants were exposed to the symmetrically averaged faces to assess their recognition of the emotion, and to the chimeric faces to determine the strength of their hemispheric lateralisation for facial emotion processing.
Before being exposed to the stimuli participants were asked to fixate on a white cross in the middle of the screen for 1000ms to ensure the child was looking directly at the facial stimuli when it appeared. This was important as the facial stimuli was only presented for 400ms. Following the presentation of each facial stimulus the participants had 10400ms to provide a response before automatically advancing to the next screen. All participants were instructed to “decide how the face is feeling and click on/point to/touch the emoji that shows that feeling”. The instruction provided differed depending on whether the child was responding using an electronic mouse, touch screen device, or by pointing and having their parent select the response for them. The latter of which was used for children who did not have access to a touch screen device, and who were not yet able to independently control an electronic mouse.
At three intervals during the chimeric face task, children were provided with the opportunity to take a break. During this break children received verbal praise and encouragement, the duration of the break was determined by the child. Administration of the chimeric face task took approximately 20-40 minutes.
Scoring. A laterality index was calculated for each child, to determine their strength of lateralisation for facial emotion processing. The laterality index was calculated by calculating the number of times the participant selected the emoticon corresponding with the emotion depicted on the right and left side of the face. The following sum was then computed for each participant 100 X (No. of right hemiface responses – No. of left hemiface responses)/(No. of right hemispace responses + No. of left hemiface responses).
Study Procedure
Ethical approval was obtained from the Lancaster University Department of Psychology Ethics Committee. Consent was received from all schools who agreed to distribute the study information to parents. Parental consent was obtained on behalf of all child participants, and oral consent was sought from the child participants during the virtual meeting.
The study was comprised of two parts, the first of which required parents to complete a series of questionnaires to provide a measure of the child’s demographic information, level of empathy, and autism severity. Parents were first presented with the demographic questionnaire to determine which additional questionnaires they were required to complete. If the parent’s responses on the demographic questionnaire indicated that their child had a diagnosis of ASD, then they were directed to, and required to complete, the EQ-Child and AQ-Child. If the parent’s response denoted that their child did not have a diagnosis of ASD they were only directed to, and required to complete, the EQ-Child. All questionnaires were completed on Gorilla (www.gorilla.sc), a cloud-based software platform for collecting data in the behavioural sciences (Anwyl-Irvine et al., 2020). All participants therefore completed the questionnaires remotely, on a personal electronic device. The questionnaires were compatible with a range of technological equipment, including a laptop, computer, electronic tablet and mobile phone.
Following successful completion of the required questionnaires, parents received an email arranging a convenient date and time for their child to complete the second part of the study, which required their child to complete the chimeric face task. The data collected during completion of the chimeric face task was linked to the parental questionnaire responses via a unique participant ID code, which was allocated to parents following confirmation of participation. Following completion of the chimeric face task, a debrief sheet and certificate was sent to the parent’s email address.
Lancaster University
07.09.2023
Ching Yee Pang
Aleeza Sulaman
Open
N/A
Data/Excel.csv
English
Data
Brooks2023
LA1 4YF
Third Parties and Police Use of Lethal Force: Evidence from the Mapping Police Violence Database
Lethal force, Third Parties, Police Citizen Interactions, Use of Force
Over recent years media coverage has highlighted the use of excessive force by some police officers. The use of lethal force towards black and other ethnic minority citizens has been identified as a cause for significant concern. Research in the bystander literature and in non-fatal force policing contexts has identified that third parties can have positive impacts in reducing the severity of these incidences. The role of third parties in fatal force events, however, has not been investigated. This is something which the current study seeks to address. The Mapping Police Violence database was used to identify a year’s worth of lethal force events in the US. Newspaper articles relating to these incidents have been coded in line with a predefined coding framework to examine the presence of third parties in these incidents, and the nature of any social relationships with third parties in relation to the type of lethal force utilised. The results revealed that third parties were present in just under half of incidences and that the presence of a third-party with a pre-existing social relationship to the citizen was associated with a lower likelihood of officers utilising forms of ‘less lethal’ force to the extent that it results in a citizen fatality. These findings highlight the potential importance of third parties in understanding the nature of lethal police citizen interactions, and also the potential protective role the presence of known others may have in reducing the likelihood of officers excessively utilising forms of less lethal force.
Sian Reid
A secondary data analysis was utilised to examine the presence of third parties in incidences of police use of lethal force. The Mapping Police Violence database (Mapping Police Violence, 2020) was the primary dataset utilised for the study. This is a freely available and open public database compiled by researchers in the US which aims to provide a record of all police involved deaths in the US. This database has been recording police involved deaths in the US since 2013, primarily gathering information through news articles published by various American news outlets. The type of force engaged in by officers that resulted in death was utilised as the outcome variable. The predictor variables were the presence of third parties, the presence of any known third parties, or unknown third parties, the number of officers present, the presence of other emergency services, the location of the incident, the race of the citizen, the gender of the citizen, the alleged presence of a weapon, the initial reason for the encounter, the presence of any digital technology capturing the event and the level of threat posed to the officer.
The Mapping Police Violence database records multiple variables in relation to these incidences, including individual and situational factors. Several of the predictor variables included in the current study have been gathered from this dataset; specifically, the type of lethal force used, the alleged presence of a weapon, the race of the citizen, the gender of the citizen, the level of threat posed to the officer, the initial encounter reason and the presence of a body worn camera. Within the current study, most of these variables have been used as recorded in the dataset, however, the level of threat posed to the officer has been recategorized. The multiple different levels of threat recorded in the dataset have been regrouped into three categories: attack (indicating the greatest level of threat to the officer), other (referring to any other level of threat), and none (for incidences in which it was clear there was no threat to the officer). In the original data only the presence of a body worn camera is recorded. For the current study this variable has been transformed to include the presence of any digital technology capturing the event, such as CCTV or smartphones, as research has found that the presence of any digital technology and not only a body camera can affect police citizen interactions (Shane et al., 2017).
The Mapping Police Violence database records the citizen’s cause of death in relation to the type of force utilised. In incidences where multiple types of force have been identified as contributing to the citizen’s death, the database records a list of all types of force involved. The types of force included in the database include gun, taser, pepper spray, baton and physical restraint. For the current study, these types of force have been grouped, to provide an outcome variable with fewer levels. The grouping of the outcome variable has been done in line with previous research looking at police use of force, which identified a gun as a distinct type of force due to the increased risk of lethal outcomes. The other types of force are grouped into a second category of other types of ‘less lethal’ force, as these types of force have been identified as alternatives to the use of a gun, which would be expected to reduce the likelihood of a citizen fatality (Sheppard & Welsh, 2022). In incidences where multiple types of force were used, the most severe form of force has been recorded; for example, if the cause of death is attributed to a gun and a taser, then this incident would be recorded as a gun as the type of lethal force utilised.
The dataset contains links to the news articles which have been used to gather information regarding each of the individual police involved death incidences. The variables included in the current study relating to the presence of others were gathered by coding these news articles which are linked in the database to the individual incidences of police involved deaths between 6th March 2022 – 6th March 2023, providing a sample of 1,257 police involved deaths. News articles are a source of information which have been identified as having certain limitations, particularly relating to potential media bias in the reporting of crime related stories (Lawrence, 2000). Research looking at the reporting of police use of force incidences by newspapers, however, has found that for many factors there was consistency between news reports and police reports of the same incidents (Ready et al., 2008). For the current study, news articles are utilised due to the promise they provide in allowing the events of police involved deaths to be examined in relation to the presence of third parties.
To identify the relevant incidences for the current study, three primary exclusion criteria were applied prior to the coding of the news articles. Firstly, to identify incidences with news articles with sufficient information to allow the presence of third parties to be examined, a minimum word count of 150 words was required in at least one of the associated news articles. Secondly, as the study’s primary interest was in the use of lethal force, which involves an on-duty officer using force, only incidences relating to on duty officers were included. Finally, incidences in which the use of force by the officer was accidental, such as car crashes that police officers were involved in, were excluded, as these events have different characteristics to those in which officers intentionally engage in the use of force towards a citizen. The application of these exclusion criteria left a sample of 1052 incidences of police use of lethal force.
To investigate the presence of others in these incidences, prior to the analysis a predefined behavioural coding scheme (Philpot et al., 2019) was created and applied to the news articles to capture the presence of third parties. This coding scheme contained 12 individual items capturing the presence of third parties and any social ties between third parties and the citizen involved in the incident (See Appendix A for the full coding scheme). Two additional items were included to capture the presence of multiple officers or other emergency services. One code regarding the location of the incident was also included to capture whether it occurred in a public, semi-public or private location. Each of the items were coded for presence with a 1, their absence recorded with a 0, or if it was not clear whether this item was present a 99 was recorded. In total 15 codes were included in this behavioural coding scheme. Here are some examples of these codes relating to the presence of third parties:
“The presence of a third-party with a pre-existing social connection to the primary citizen involved”
“The presence of more than one officer”
“The presence of a third-party with no pre-existing social connection to the primary citizen involved”
To facilitate the process of coding the news articles in line with the coding scheme, a Qualtrics survey (https://www.qualtrics.com) was created. This survey presented the individual items within the coding framework in a questionnaire format, allowing the items to be coded in the format of closed ended responses to questions relating to the presence of third parties. The responses from the survey were then transferred to an Excel document to allow the data to be prepared for analysis.
Ethical approval has been obtained for this study. The study has been reviewed and approved by a member of the Lancaster University Psychology Department, the ethics partner of the supervisors.
The reliability of the coding scheme and its application to the news articles was assessed through the double coding of 10% of the sample by a second researcher separately to the primary researcher. To assess the level of agreement between the two researchers for each variable, Gwet’s AC1 (Gwets, 2014) coefficient was calculated. In line with the recommendations of Landis and Koch (1977), the resulting coefficients were interpreted in the following way: a value of 0.4 or above indicating moderate agreement, a value of 0.6 or above indicating substantial agreement, and finally a value of 0.8 or above, indicating almost perfect agreement between raters’ scores. For 13 of the variables an agreement level of substantial or almost perfect was reached, as seen in table 1 (appendix B). For the variable relating to the third-party being a friend of the citizen there was no variation in responses (i.e., 100% agreement), and therefore a coefficient could not be calculated. For the location variable, only a moderate level of agreement was found, as a result this variable was excluded for the purpose of analysis.
Figure 1 depicts a flowchart of the process undertaken to sample the relevant incidences. The first part of the flowchart shows the initial process that was undertaken to identify all police involved deaths recorded in the Mapping Police Violence database in the prior 12 months. Following the initial data collection procedure descriptive statistics were run which highlighted that in the initial sample of 1052 incidences there was very limited variation in the outcome variable of the type of lethal force utilised by officers, with 990 incidences involving a gun as the primary cause of death, and only 62 incidences involving other forms of force. In this initial sample a citizen’s cause of death not involving a gun would statistically be considered a rare event, which would have presented challenges in utilising this variable as the outcome in any subsequent analyses. In line with the recommendations of research (Shaer et al., 2019), an oversampling approach was chosen to overcome the limitations of having a rare event in the outcome variable, with further incidences in the dataset that did not involve a gun as the cause of death being oversampled so at least 10% of the sample involved a cause of death other than a gun. As can be seen in figure 1, for these incidences to be as similar to the primary sample as possible, they were only sampled for the three preceding years to limit any additional sample variation that may have been introduced by sampling a wider date range. This led to the identification of a further 182 incidences where the citizen’s cause of death did not involve a gun. The same exclusion criteria were then applied to this sample, with a further 65 incidences excluded, leaving a sample of 117 additional incidences which were coded in line with the same procedure as the initial sample. This oversampling procedure led to a final sample of 1169 incidences.
The data analysis involved chi square tests of independence, to examine whether the presence of others during fatal police citizen interactions had a statistically significant relationship with the outcome variable of the type of lethal force utilised by officers. Due to the exploratory nature of the study there was not a predicted direction or nature of the relationship between the predictor variables relating to third-party presence and the type of fatal force utilised by officers (McIntosh, 2017). Prior to the main analyses, descriptive statistics were run to investigate distributions within variables and to allow any rare event variables to be identified.
6th September 2023
Charlotte Thompson
Open
N/A
Excel.csv
r_file. R
English
Data
Reid, 2023
Investigating infant expectation on object search tasks
Chi-squared, Correlation, Factor analysis, Linear mixed effects modelling
The current study aims to distinguish between Piaget’s (1954) theory of object understanding, highlighting the role of object permanence on A not B task performance, and Diamond’s (1985) theory highlighting the role of motor demands and lack of ability to inhibit habitual behaviours during the task. These two theories differ in their predictions for the expectations of the infants taking part, with Piaget (1954) predicting that infants’ lack of object permanence causes poor performance on the task and Diamond (1985) predicting that infants understand the movement of objects and a lack of inhibition of habitual behaviours cause error in performance. We tested 15 nine-month-old infants on a looking version of the A not B task. The use of impossible and possible outcomes was also incorporated on B trials, with the object being revealed from either the correct or incorrect location (e.g., see Ahmed & Ruffman, 1998). Infant first look direction, accumulated looking time during trials and the number of social looks initiated post-outcome, were used as measures. We found significant evidence of the ‘AB’ error during trials, with a significantly increased number of incorrect first looks on B trials. There was also a descriptive pattern showing surprise at object location reveals with increased number of social looks during B compared to A trials, though this was not significant. Accumulated looking analysis showed that infants looked longer on A than B trials, suggesting that infants expected the object to be in location B on B trials, demonstrating infants’ ability to understand objects and supporting Diamond’s (1985) theory. However, implications for a small sample size and presence of individual differences on interpretation of looking time data are discussed. Implications in theory and future research are suggested and overall, results provide support for the application of Piaget’s (1954) theory and suggest that infants have limited object understanding based on their displayed expectations during testing.
Leah Murphy
3.1. Participants
In this study, 15 participants took part, aged 8 months and 12 days to 9 months and 27 days old (M = 9 months and 3 days, SD= 11.3 days). Six further infants were excluded from data analysis as they became too fussy to complete the study. Participants were recruited from the Lancaster Baby lab database, along with the Lancaster Baby lab Facebook page and were also recruited via word of mouth from guardians taking part in the
study.
3.2. Materials
The video stimuli were created using Canva software (Canva.com, 2023) and was uploaded onto ‘Habit 2’ software (see Oakes et al., 2019) to display the stimuli during testing and to measure the accumulated looking time of the infant participants. The stimuli involved a novel object obtained from the NOUN database (Horst & Hout, 2016). A camera was used to record the social looks exchanged between the infant and guardian, as well
as the direction of the infants’ first looks during testing.
3.3. Design
This study had a within-subjects design, with all participants being exposed to the same experimental conditions and the same stimuli. To counterbalance for location effects, half of the participants witnessed A trials being hidden in the box on the left, whilst the other half witnessed the object being hidden in the box on the right during A trials. The presentation of the accurate and inaccurate B trials was further counterbalanced across participants, as half of the participants viewed the inaccurate B trials first, and the other half viewed the accurate B trials first.
3.4. Ethical approval
Ethical approval for this study was granted by the departmental ethics committee (DEC) at Lancaster University. Guardians were recruited via their preferred contact method and were sent the participant information sheet to read before agreeing to take part in the study. A date and time of testing was arranged at the Babylab building at Lancaster University, via telephone or email. Upon arrival, guardians were presented with the consent form to sign and initial all points before being allowed to take part. They were also given the opportunity to ask any questions about the study and were informed that they could withdraw at any time.
After the study, the guardian received a five-pound contribution to travel costs, along with a free children’s book for the infant, as a reward for taking part in the study. The guardian also received a debrief sheet to read and to take home, providing them with all contact information of the lead researcher, if they wished to ask any questions or to withdraw from the study.
3.5. Procedure
The testing took place in a private room within the Whewell building at Lancaster University. The infant and guardian were sat in front of a computer screen with the infant sat in a highchair positioned directly in front of the screen, and the guardian sat in a chair to the side, slightly behind the infant (to allow researchers to see clearly when the infant initiated a social look). The experimenter sat behind a divider at a computer, out of sight of the infant and guardian. A social engagement video of the experimenter saying, “Let’s hide the blap, can you find the blap?” was presented to the infants at the start of the experiment and between each trial, to insert social communication and guide the attention of the infant to the screen before the stimuli were presented. The infant then watched a series of video stimuli in which a novel object appeared on the screen and moved into one of two boxes, both boxes were then covered (the object was hidden), and a there was a delay period of five seconds (see figure 1). After the delay period, both boxes were revealed, and the location of the toy was visible to the infant. Any movement of the object was accompanied by a sound to guide the attention of the infant to the object, but this sound was not present when the object was revealed to avoid any leading factors when measuring infant expectation. Instead, the occluders made a simple “whoosh” sound when they were removed, to ensure the infant was paying attention. After five identical A trials, the object was then hidden in the second location and the process was repeated consisting of six B trials. However, during the B trials, the object was hidden in the second location, but was either revealed to be in the correct (accurate) or incorrect (inaccurate) location (see figure 2). This variation in outcome was presented alternately to the infant, with the object being revealed from the incorrect location for three out of the six B trials. The study lasted for approximately 10 minutes per participant.
3.6. Behavioural coding
Infant looking time was coded online as trial lengths were infant controlled. Each trial ended when the infant looked away for four seconds. As this controlled the trial length, this was not double coded as this inherently will lead to a high agreement level. For the coding of infant first look and number of social looks, the videos recorded of the participants were saved and uploaded onto Microsoft OneDrive to be offline coded. First look was defined as the direction that the infant first looked towards once the occluder was removed and the object was revealed. On trials where the infant was not looking as the occluder was removed, the first look was defined as the direction in which they looked once their gaze returned to the screen. The first look direction was coded as correct and incorrect. The number of social looks initiated by the infant per trial was also measured during coding, defined by the infant turning towards the guardian during each trial after an outcome was revealed. Twenty percent of the videos were dual coded and there were no discrepancies between researchers during the dual coding process for first looks (r = 1, p<0.01) or social looking (r= 1, p<0.01).
2023
Shiyu Pang
Yuewen Qin
open
none
.xslx
English
dataset
Murphy(2023)
The Effects of Posture on Body Part Width Representations
Body perception, affordances, somatosensation, visual perception
Despite the ubiquity of our bodily experiences, our representations of our body’s size are not geometrically accurate. For example, when estimating the length of body parts using the hand as a metric, consistent patterns of distortions across body parts are observed. Given the presence of these distortions, some have proposed that representations of length and width emerge directly, or indirectly, from the organisation of somatotopic maps in somatosensory cortex, rather than from their actual relative dimensions. However, whilst length representations are well researched with respect to this notion, less is known about representations of body part width across the body. Moreover, it is unclear from previous research whether body part width representations may be confounded by participants’ posture. Specifically, individuals have shown an enhanced tendency to overestimate body part width when seated upon a chair, suggesting that the chair may become incorporated into the body representation. Consequently, the aim of the current investigation was to further elucidate how width is represented across body parts and whether posture moderates these representations. Participants estimated how many hands widths made up the width of the back, shoulders, hips, torso, thigh, and head in one of three conditions: standing (n = 37), seated upon a chair (n = 33), or seated upon a backless stool (n = 39). Whilst estimates did differ across body parts, no effect of posture was observed. Moreover, the patterns of distortions observed differed from those seen in previous investigations. Results therefore indicate that body part width representations are neither accurate nor fixed, rather, they show distortions which vary across individuals and contexts. It is proposed that inter-individual heterogeneity in width representations may result from humans possessing alternative perceptual mechanisms for judging aperture passability. Therefore, maintaining fixed width representations is unnecessary, and hence too energetically costly to maintain.
Lettie Wareing
Method
Participants
Ethical approval for this study was obtained from Lancaster University Psychology Department on 31st May 2023.
As this study aimed to investigate body part width representations in healthy populations, only participants aged 18-55 years without any physical, or mental impairment were included in the study. However, as previous research (Readman et al., 2021) using the same paradigm for length estimates has shown no influence of anxiety or depression on body part estimates, participants with diagnoses of these conditions were not excluded. Participants were excluded if they had any current or historic diagnosis of cognitive impairment, as this can affect instruction comprehension (Han et al., 2011), or visual impairment, to ensure difficulties in seeing the body parts did not confound findings. Furthermore, given the associations between other psychiatric impairments (e.g., Priebe & Röhricht, 2001), neurological impairments (e.g., Blanke et al., 2004), or eating disorders (Mölbert et al., 2017) with distorted body perceptions, individuals with a current or historic diagnosis of a condition falling within any of these categories were excluded.
A total of 123 (61 females) participants ranging from 18 to 68 years (M = 28.80 years, SD = 10.79) were recruited via opportunity sampling for this study. Participant recruitment was ended before the required N = 150 due to time constraints. All participants were entered into a draw to win one of two £25 vouchers as an expression of goodwill. A total of 15 participants were excluded for failing to meet the inclusion criteria, leaving a final sample of N = 108 (50 females). Participants were aged 18 to 55 years (M = 27.98 years, SD = 9.56); the majority of participants were right-handed (n = 99) and over half the participants had normal vision (52.78%), with the remaining participants having corrected-to-normal vision.
Reasons for exclusion included a current or historic psychiatric impairment (n = 2) or eating disorder (n = 4), falling outside the study age restrictions (n = 3), visual impairment (n = 2), being pregnant (n = 1), failing to provide demographic information needed to determine eligibility (n = 2), and a self-reported misunderstanding of task instructions (n = 1).
Design
This study constituted a 3x6 mixed design with condition (standing, chair, or stool) as the between-subjects variable and body part (torso, hips, shoulders, back, thigh, or head) as the within-subjects variable. The dependent variable was participants’ accuracy ratios for each body part (actual size/ estimated size) where an accuracy ratio of over 1.0 indicated overestimation, and under 1.0 indicated underestimation of body part width.
Materials and Procedure
After providing their consent, participants completed a self-report demographic and clinical questionnaire administered via Qualtrics (Qualtrics, Provo, UT) which asked about participants’ age, biological sex, preferred hand, and details regarding their neurological, cognitive, and psychiatric history.
Following this, participants were randomised to one of the three conditions (Standing, Chair, or Stool). In each condition, participants were asked to estimate how many hand widths of their dominant hand made up the width of six different body parts: the torso, shoulders, hips, back, head, and thigh. Participants were instructed to be as accurate as possible, using fractions where necessary. They were asked to refrain from touching the body part with their hand, or basing estimates off estimates for previous body parts if the two body parts were proportionally related. The researcher defined each body part verbally and pointed to their endpoints on their own body prior to the participant making their estimate.
Participants in the standing condition performed all estimates whilst stood upright, without leaning on any surfaces. In the chair condition, participants were seated upon a standard desk chair with a high back and no arm rests. In the stool condition, participants were seated upon a fixed height bar stool with no back. The condition completed by participants was counterbalanced, and the order of body parts estimated was randomised.
After making their estimates, the researcher used a soft tape measure to measure the actual width of the cued body parts before debriefing participants. The study took around 10 minutes to complete.
Analysis
Prior to conducting the analysis, outliers were removed using the median absolute deviation (MAD) approach. This procedure involves removing participants whose accuracy ratios deviated more than three absolute deviations from the median for a given body part. The MAD approach was chosen as it is more robust than traditional methods of outlier detection based upon standard deviations from the mean (Jones, 2019; Leys et al., 2013).
To calculate the dependent variable of accuracy ratios, first, participants’ hand estimates for each body part were converted to centimetres by multiplying their estimate in hands by their measured hand width. After this, estimates for each body part were divided by the actual width of the body part to produce an accuracy ratio.
To test the study hypotheses, data was analysed using a 3x6 mixed ANOVA using the rstatix package available from RStudio (Version 4.2.1). Body Part was entered as the within-subjects variable, and Condition as the between-subjects variable. The assumption of normality was checked using the Shapiro-Wilks test, and the sphericity assumption via Mauchly’s test. Partial eta-squared was used as a measure of effect size.
Though frequently used in analysis, frequentist statistics are not without limitations. It is typically assumed that a p-value of <.05 is evidence for the alternative hypothesis, however this value only represents the probability of obtaining results as extreme as those observed, if the null is true (Wagenmakers et al., 2018). Therefore, data which is unusual under the null hypothesis is not automatically any less unusual under the experimental hypothesis (Wagenmakers et al., 2017). Moreover, a non-significant finding in frequentist analyses cannot be taken as evidence in favour of the null hypothesis (Kruschke & Liddell, 2018). In this regard, Bayesian statistics have several advantages over frequentist statistics including the ability to incorporate prior knowledge, quantify the degree of uncertainty surrounding the existence of an effect, and the ability to quantify the strength of evidence in favour of the null, or alternative hypotheses (see Wagenmaker et al., 2018 for a discussion).
Consequently, to provide further support for conclusions drawn using frequentist analyses, a Bayesian Mixed ANOVA was conducted using the anovaBF function from the BayesFactor available in RStudio (Version 4.2.1). Default priors were used given that these reflect average effect sizes observed across all psychological experiments, and hence are likely to be more reliable than priors drawn from a single, potentially methodologically flawed, study (Rouder et al., 2012).
Where a significant main effect of Body Part or Condition was observed, Holm-Bonferroni adjusted frequentist, and Bayes Factor, pairwise t-test comparisons were conducted to determine the pattern of differences underlying these effects.
In addition, to determine whether body part width estimates differed significantly from 1.0 (i.e., an unbiased estimate), Holm-Bonferroni adjusted frequentist, and Bayes Factor, one-sample t-tests were conducted for each body part.
To judge the strength of evidence provided by the Bayes Factor analyses, Kass and Raftery (1993) criteria was used. By this criteria, Anecdotal evidence is regarded as inconclusive. Percentage error (a measure of certainty in the estimate) was reported alongside Bayes Factors, where <20% is regarded as an acceptable level of uncertainty (Van Doorn et al., 2021).
Lancaster University
2023
Leanna Keeble
Open
None
Data/Excel.csv
English
Data
Wareing2023
LA1 4YF
Inner Speech and Its Role in Purchasing Decision-Making Process: Analysis of Within-Subjects Experiment and Questionnaires
inner speech, purchasing behaviour, memory, decision-making.
Inner speech is a cognitive function related to language processes. Based on its functions reflecting information processing and memorising, it may link to the purchasing process, which includes searching and evaluating product information. Inner speech may also help people think and imagine using the product in the future during their purchasing process.
This study discussed and investigated the role of inner speech in the purchasing process and how it might affect the decision-making time. This study also mentioned how inner speech may be identified and suppressed. Participants’ data was collected through experiments and several questionnaires. The findings indicated that inner speech might help people in Information Search and Alternative evaluation and affect decision time. The findings also suggested what people may consider and how they use inner speech.
By uncovering the potential relationship between the purchasing process and inner speech, this research provided valuable information for marketing and psychology research fields. It gave companies some suggestions for practical use, reflecting how people may use inner speech during the purchasing process.
Han-Yi Wang
Methods Section:
This study was approved by ethics committees at Lancaster University. There were no ethical issues for researchers managing the personal information. The participants’ information remained anonymous and were assigned subject ID (P01, P02, P03…, P30 in Experiment 1 and PCT01, PCT02, PCT03…, PCT30 in Experiment 2). All data were stored anonymously with no identifiable information.
Participants were given the Participant Information Sheet (PIS) before participating in the experiments. On the day of testing, they asked any questions they might have, then consented to attend the experiment in person or via online platforms like Microsoft Teams, Zoom, or Google Meet to ensure that the suppression was active when needed. The experiment took approximately 30 minutes, including answering all questionnaires. The experiment was held in the participant’s home or a place where no one spoke so that the participant would not be disturbed by any chance.
Experiment 1
Participants
G*power suggested 52 participants within groups using t-tests and multiple mixed linear regression models, with a .4 effect size and .05 (5%) a-error probability in 80% power (1-b error of probability) (Brysbaert, 2019). Thirty participants were recruited in this experiment with no record or history of neurophysiological disorders, such as dyslexia or aphasia, to ensure that no conditions influence the result and affect the participant to complete the tasks in the experiment. The recruitment process included in-person invitations around campus and social media messages to reach diverse participants.
Although only 30 participants were recruited in this experiment, the results of the t-tests suggest that the effect size (see Experiment 1 result section) may be enough for testing the hypothesis.
Design
This study was an experimental within-subjects design. Participants simulated purchase experience in the suppression task and the control task without interference assigned to them. The independent variables were self-rating agreements on information search and alternative evaluation and participants’ average decision time in the suppression and control tasks. The dependent variables were inner speech frequency in five dimensions measured by the Inner Speech Frequency Questionnaire (VISQ).
Quantitative data were analysed using R to conduct t-tests, GLMM and CLMM. Secondly, qualitative data were collected through questionnaires and categorised into different variables to identify why participants made the decisions and their inner speech content during the purchasing process.
Overall, the experiment aims to investigate how people use inner speech during purchasing and whether Articulate Suppression task and task without interference influenced decision time and agreement score on information search and alternative evaluation.
Materials
Stimuli
Participants viewed six product sets (stimuli), which information was copied from the official website. To prevent participants from focusing on the effect of the products’ brands and prices (Albari & Safitri, 2020), the products in each set were the same brand with similar or the same price, unisex, and recognisable, although these products might not exist or remain the latest information on the market.
Two-item Statement Questions (see Appendix B)
Participants rated the two statements on a seven-point Likert score from strongly disagree to strongly agree (Maity & Dass, 2014) to identify the Information Search and Alternative evaluation agreement level between tasks. Then, participants were asked: “Which product did you choose? Why?” after each purchasing decision.
Variety of Inner Speech Frequency Questionnaire (VISQ, see Appendix C)
The Inner Speech Frequency Questionnaire (Alderson-Day et al., 2018) included twenty questions asking participants to generally rate their inner speech frequency after the mock e-commerce purchasing tasks with a 7-point Likert scale ranging from "Never" to "All the time". Questions 7 and 15 were reversely coded; the value should be reversely calculated when doing analysis.
Experiment 1 Qualitative Questions (ExpQ1, see Appendix D)
After participants finished all the tasks (six decisions), they were asked to answer three questions at the end of the experiment. These questions gathered qualitative data about the participants’ experiences during the mock e-commerce purchasing tasks and what they had in mind.
Procedure
Figure 2 illustrates the diagram of Experiment 1. Participants were invited and consented to join the research to do Suppression and Control (without interference) tasks.
Each task contained three product sets; participants were asked to imagine and choose a product for themselves or a friend according to the provided information on the mock e-commerce channel (Maity & Dass, 2014). The screen of the researcher or participants presented the information, including the price and details of the product set. Since these two tasks are counterbalanced and randomly ordered, participants repeated the decision-making process three times in the control task and the other three in the suppression task. After each decision, participants answered the two-statement questionnaire and explained which products they chose and why they chose them. According to different tasks, they started with the control task by themselves. However, they were asked to practise counting out loud from 1 to 4 following 160 bpm metronome sounds until the researcher ensured they remained suppressed before starting the suppression task.
Then, they answered VISQ, which measured their inner speech frequency and qualitative questionnaires (ExpQ1) to understand how they used inner speech when viewing the products in the last part of the study.
Analysis
R was used to analyse the quantitative data to identify the task differences via t-tests and the relationship between variables in two tasks via Generalised Linear Mixed Effect Models (GLMM) and Cumulative Link Mixed Model (CLMM). When conducting the GLMM with family gamma, the quantitative data will follow the standard procedure of data trimming and keep the trimmed data within 5% or 2.5 standard deviations (Berger & Kiefer, 2021).
The qualitative coding scheme (See Appendix F) was created to identify what participants considered and what they said to themselves using inner speech during the experiment. The coding process involved re-reading the data to identify and assign relevant contexts to the appropriate categories. For example, if participants mention that they have used the product before, the value of the variable “Memory” increases by one unit. These variables were then calculated to identify what factors influenced participants’ purchasing decision-making more. Following the same coding scheme, what kind of inner speech was used when viewing the products could also be found. For example, people may ask themselves questions or repeat the product in mind.
In summary, Quantitative and qualitative data were analysed to report the results for different purposes and test the hypothesis in this research.
Experiment Optimising
The task without interference in Experiment 1 may not be a reasonable control task since it might include the secretary task effect, as participants were asked to do both tasks and be influenced after they did the suppression task when they were doing the control task.
As a secretary task, the finger-tapping task, which has been used in inner speech experiments, could be the better control task in Experiment 2 (Emerson & Miyake, 2003; Wallace et al., 2009). Although Finger-tapping might influence working memory’s function and influence people to memorise (Armson et al., 2019; Kane & Engle, 2000; Moscovitch, 1994; Rose et al., 2009), Rogalsky et al. (2008) also mentioned that the performance of people’s understanding of complex sentences might decrease but not as much as suppression occur.
Therefore, doing the second experiment was motivated to replicate the results with a better control condition involving Finger-tapping.
Experiment 2
Participants
Based on the findings of Experiment 1, another 30 participants were recruited with the duplicate requirements as the first experiment. The recruitment requirement and process were the same as in the previous experiment.
Design
The independent variables were similar to Experiment 1, while the only difference was that the control task here had been changed into the Finger-tapping task. The goal of the whole design is to replicate the results of Experiment 1 to investigate the role of inner speech in the purchasing process.
Materials
Experiment 2 applied the same materials used in Experiment 1. The only difference was the qualitative questions after tasks. In Experiment 1, participants answered “Experiment 1 Qualitative Questions” at the end of the experiment. However, to better understand the difference between tasks, they were asked to answer a similar questionnaire (see below) after each task to discover the inner speech used in the two tasks.
Experiment 2 Qualitative Questions (ExpQ2, see Appendix E)
Participants were asked to answer three questions about their experiences during the mock e-commerce purchasing tasks and what they had in mind for the Suppression and Finger-tapping tasks separately.
Procedure
The procedure was the same as the first experiment, except for adjusting the control task and the order of the qualitative questionnaire (ExpQ2). Figure 3 illustrates that participants were invited to the experiment using the same stimuli, similar questionnaires, and the same method of presenting stimuli (participants joined in person or via online platforms) with Suppression and Finger-tapping tasks. Participants were asked to practice counting 1,2,3,4 out loud or tapping their index, middle, ring, and little fingers in order (see which task came first) following metronome beats at 160 bpm before the researchers decided to move on. They were asked to view the product set by imagining choosing one for a friend or themselves three times in each task. Participants answered two statements and answered what product was chosen and why after each decision they made. Then, they were asked to answer three Qualitative questions (Appendix E) after each task. They repeated another task in the same process afterwards with a 2-minute break between tasks. After they finished the Finger-tapping and Suppression tasks, they answered VISQ questions at the end of the experiment.
Analysis
R was also used to analyse the quantitative data for the same purposes and followed the same data-trimming procedure if needed. The same coding scheme was followed to generate the result that could replicate and optimise the clarity of the Experiment 1 results. Overall, the second experiment is to generate the same or more evident results as Experiment 1 and to find more valuable information for the different inner speech used between tasks.
In conclusion, these two experiments and the analysis might give this research a deeper understanding of inner speech and its role and provide more precise information on how inner speech may related to the purchasing process.
Lancaster University
2022-23
Melanie Thomas
Vickie Huang
Open
None
The data set is in csv format.
English
Data
Wang2023
LA1 4YF