Lights, Camera, Action: Investigating Advertisement Susceptibility in Films Amongst Individuals with Parkinson’s Disease and Controls.

Lights, Camera, Action: Investigating Advertisement Susceptibility in Films Amongst Individuals with Parkinson’s Disease and Controls.

Elena Ball

07.09.2022

Product placement is the merging of entertainment with advertising, and its presence in our daily lives is increasing. Despite this, there is an inherent lack of consideration of its influence amongst vulnerable populations such as individuals with Parkinson’s disease (PD). Research suggests that individuals with PD have reduced inhibitory control (IC) which may drive impulsive behaviours. A concernment, therefore, is the influence that product placement may have on the purchase behaviour of individuals with PD alongside a possible propensity to partake in risky and impulsive behaviours. Thus, this study aimed to examine whether reduced IC increases the likelihood that an individual with PD will be susceptible to product placement. The study adopted an experimental approach, recruiting 20 healthy younger controls, 20 healthy older controls, and 13 individuals with mild to moderate PD to participate in watching two films containing product placement; one featuring Coca Cola and the other an Audi. A pre and post product placement questionnaire was used to measure change in purchase behaviour before and after exposure to product placement, and an antisaccade eye tracking task and a Stroop task was used to measure IC. An ANOVA indicated that IC was significantly impaired in individuals with PD compared to healthy controls. Despite this, linear mixed effects modelling suggested that IC may not be a factor that increases the likelihood that an individual will be more susceptible to product placement. Implications of these findings are discussed relative to other clinically vulnerable populations with similar cognitive impairment symptomology, and the consequent need for future research to continue to explore product placement susceptibility amongst vulnerable populations.

Parkinson’s Disease, Inhibitory Control, Product Placement Susceptibility

Method
Participants
A voluntary sample of 54 participants were recruited, 20 healthy younger controls (YC) (16 females and four males, (Mage= 22.70, SDage= 2.42)), 20 healthy older controls of comparable age to those with Parkinson’s (OC) (females and males, (Mage= 66.85, SDage= 8.53)), and 15 adults with mild-moderate idiopathic PD (females and males, (Mage= 65.00, SDage= 7.84)). As this research area is entirely novel this sample size was modelled on comparable population studies that have explored IC (Meyer et al., 2020; Paz-Alonso et al., 2020). YC were defined as young adults aged between 18 to 26 years old with no neurological or cognitive conditions (Stroud et al., 2015). OC were defined as adults aged between 50 to 85 years old with no neurological or cognitive conditions (Zhang et al., 2020). The participants with PD had been diagnosed with mild-moderate idiopathic PD, characterised by mild-moderate impairments of motor and cognitive functioning (DeMaagd & Philip, 2015).
The exclusion criteria for both the healthy controls and individuals with PD were those who had a diagnosis of any additional neurological or cognitive conditions other than PD. Moreover, given that visual impairments may affect the visual experience of product placement, all participants were screened for red-green colour blindness using the Ishihara test. The standardised cut off for normal vision is 15 (Rodriguez-Carmona & Barbur, 2017), therefore, participants who score 14 or less were excluded as this is indicative of the presence of red-green colour blindness.
All participants had normal or corrected-to-normal vision. The Addenbrooke’s Cognitive Examination-III (ACE) was used to screen for the presence of cognitive impairment (Bruno & Vignaga, 2019). Participants’ data was only included in analysis if participants achieved a score within the normal range (≥ 82 out of 100). Following this exclusion criteria, one PD participant’s data was removed. Research has shown saccadic eye movements to be influenced by cognitive dysfunction (Hutton, 2008; MacAskill et al., 2012), thus cognitive impairments need to be screened for as this study is measuring saccadic eye movements as a measure of IC. Subsequently, following exclusion criteria, 53 participants’ data was included within analysis.
PD participants were selected who were at a Hoehn and Yahr Stage three or less (see Table 1 for background characteristics for participants attached in the files below). The Hoehn and Yahr is used to give a summary of the laterality and severity of PD symptomology (Readman et al., 2021b). Five participants presented unilateral symptoms only (stage one), seven participants presented bilateral symptoms with no impairment of balance (stage two) and one participant presented bilateral symptoms with some postural instability but were not physically dependent (stage three). PD symptomology was assessed using the Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) (Evers et al., 2019). All PD participants were tested under their usual medication regimes and were in a typical functioning ‘ON’ phase. Eight participants were taking a dopamine agonist (e.g., Ropinirole), eight participants were taking a combination drug (e.g., Madopar), six participants were taking a monoamine oxidase inhibitor (e.g., Rasagiline), and two participants were taking a Catechol-O-Methyl Transferase (e.g., Entacapone).
YC were recruited through the researcher’s social network. Whereas both OC and individuals with PD were recruited established research interest databases (OC C4AR database; PD MRR PD interest database (FST2005)).
Materials
Health and Demographic Questionnaire
The health and demographic questionnaire (HADQ) was developed and distributed using Qualtrics (Qualtrics, 2022), an online software that aids the process of building, distributing, and analysing surveys (Carpenter et al., 2019). The HADQ was comprised of four distinct subsections pertaining to both the participants general demographics, and more specific health related measures.
Demographic Questions. For participant group allocation, participants were asked for their age, sex, and whether they held a diagnosis of PD. Information about participants’ age also afforded the opportunity for exploration into the possible effect of age as well as PD on product placement susceptibility.
The Hospital Anxiety and Depression Scale (HADS). The HADS is a 14 item (7 items pertaining to anxiety and 7 items pertaining to depression) self-report assessment of anxiety and depression suitable for both psychiatric and non-psychiatric populations (Stern, 2014). All items are rated on a 4-point severity scale with a total score of 11 or more being indicative of probable anxiety and depression respectively (Caci et al., 2003; Edelstein et al., 2010). Literature has found HADS to be high in construct validity and very good internal consistency was observed when measuring anxiety (Cronbach’s α = .83) and depression (Cronbach’s α = .82) (Bjelland et al., 2002; Johnston et al., 2000; Mondolo et al., 2006).
Edinburgh Handedness Inventory. The Edinburgh Handedness Inventory is a 10-item self-report questionnaire in which participants are asked to indicate a preference for which hand they would use when completing a range of daily activities (e.g., brushing teeth) (Robinson, 2013). Through this a handedness score ranging from 100 (strong right) to -100 (strong left) deduced. Excellent internal consistency was observed in the 10-item Edinburgh Handedness Inventory (Cronbach’s α = .94) (Fazio et al., 2013). Previous literature suggests that handedness and eye-dominance are correlated because of hemispheric specialisation (McManus,1999; Willems et al., 2010), therefore establishing participants’ handedness was indicative of their dominant eye when measuring IC through saccadic eye movements.
PD Diagnosis questions. Participants with PD were asked to provide specifics relating to their diagnosis, including years since diagnosis, years since presumed onset, and what medication, and its dosage, they are prescribed. These items were necessary to investigate whether PD severity and medication type influence product placement susceptibility.
Screening Assessments
Cognitive Impairments. The Addenbrooke’s Cognitive Examination-III (ACE) is a cognitive assessment that screens for the probable presence of cognitive impairments (Noone, 2015). The ACE is comprised of 24 items that analyse attention, memory, fluency, language, and visuospatial processing (Bruno & Vignaga, 2019). Very good internal consistency was observed in the ACE (Cronbach’s α = .88) (Kan et al., 2019) and validity (Matias-Guiu et al., 2017; Takenoshita et al., 2019).
Visual Impairments. The Ishihara test is a reliable (Birch, 1997) 17 item assessment for red-green colour blindness that requires participants to read aloud a set of numbers on Ishihara plates that are made up of coloured dots (Marey et al., 2015).
PD Symptomology. MDS-UPDRS is a tool to measure the progression of PD symptomology (Evers et al., 2019). MDS-UPDRS is comprised of a series of tasks that assesses PD symptomology within the last week, in the domains of mentation, behaviour and mood, activities of daily life, motor abilities, and complications of therapy (Holden et al., 2018). Very good internal consistency was observed in the MDS-UPDRS (Cronbach’s α = .90) (Abdolahi et al., 2013) and valid assessment of PD symptomology severity (Goetz et al., 2008; Metman et al., 2004).
Measures of Inhibitory Control
Eye Tracking Tasks. The prosaccade and antisaccade tasks were created using Experiment Builder Software Version 1.10.1630 and the data was extracted and analysed using Data Viewer Software. Eye movements were recorded via the EyeLink Desktop 1000 at 500 Hz. Whilst recording eye movements, participants were asked to place their chin on a chin rest to reduce their head movements. Participants sat approximately 55cm away from the computer monitor (monitor run at 60Hz).
Firstly, participants were asked to complete the 4-point calibration task to improve eye tracking accuracy (Pi & Shi, 2019). In this task participants were asked to follow a red target around the screen as it moved up, down, left, and right. Next, participants completed the prosaccade eye tracking task. To centralise participants’ gaze, participants were instructed to look at a white fixation target displayed on a computer screen for 1000ms. Participants were then instructed to look towards a red lateralised target that appeared on screen for 1200ms at a 4o visual angle either to the left or to the right of where the white central dot had been located, as quickly and as accurately as possible (Readman et al., 2021a). The eye tracking equipment measured participants’ saccades and latencies (how long it took for participants to fixate on the red target). A total of 16 gap trials were presented with a blank interval screen displayed for 200ms between the extinguishment of the white fixation target and the initial appearance of the red target, which resulted in a temporal gap in stimuli presentation. The prosaccade task was incorporated to ensure that alternations in participants antisaccade task performance were not due to impaired prosaccades and rather are indicative of alterations in IC.
For the antisaccade task, participants were first asked to look at a central white fixation dot for 1000ms to centralise their gaze. Participants were then asked to direct their gaze and attention focus to the opposite side of the screen to where a green lateralised target was presented for 2000ms at a 4o visual angle either to the left or to the right of where the white central dot had been located, as quickly and accurately as possible (Derakshan et al., 2009). See figure 1 above for a visual display of an antisaccade task. The eye tracking equipment measured participants’ saccades, latencies (how long it took participants to fixate their gaze to the opposite direction to the green target), and error rates (how many time participants incorrectly looked at the green target). A total of 16 gap trials were presented with a blank interval screen displayed for 200ms between the extinguishment of the white fixation target and the initial appearance of the red target, which resulted in a temporal gap in stimuli presentation.
Stroop Test. The Stroop test was conducted using PsyToolkit’s free online demonstration (PsyToolkit, 2022). Unlike in the original Stroop test whereby participants had to say the ink colour aloud (Stroop, 1935), using PsyToolkit’s online Stroop test allowed for a more accurate measurement of participant’s reaction time (ms) through pressing the key corresponding to the ink colour (Brenner & Smeets, 2018). Participants completed the Stroop test on a HP ProBook 470 G5 17.3” laptop (HP, 2022), and were sat approximately 30cm away from the laptop. Presenting the Stroop test on this laptop enabled participants to view the test on a large screen, thus improving the accessibility of the test. The colour words presented to participants were ‘red’, ‘green’, ‘yellow’, and ‘blue’.
Participants were instructed to press the key corresponding to the initial letter of the ink colour of the printed word presented on screen as quickly and accurately as possible. For example, the correct answer for RED would be if the participant pressed the key ‘B’ for blue. A total of 40 gap trials were presented. For each trial, a colour word was presented on screen for 2000ms. The colour word was either congruent (the colour word and the meaning are the same, e.g., GREEN) or incongruent (the colour word and the meaning is different, e.g., GREEN). There was a 100ms gap in presentation of the word in which a white cross was presented on a black interval screen. Participants’ congruent and incongruent reaction times (ms), correct Stroop score (correctly identified ink colour out of 40), and Stroop effect (incongruent reaction time (ms) minus congruent reaction time (ms)) were recorded.
The ease at which the Stroop test can be conducted in a non-laboratory environment and the simplicity at which the colour words can be translated into other languages, increases its accessibility and universality as a measure of IC (Gass et al., 2013). This assessment would, however, be an invalid measure of IC for individuals affected by colour blindness or dyslexia, limiting the populations the Stroop task can assess (Scarpina & Tagini, 2017).
Product Placement Film Clips
The incorporation of film clips containing product placement was guided by the prominent use of film clips within previous research that had investigated product placement susceptibility (Kamleitner & Jyote, 2013; Yang & Roskos-Ewoldsen, 2007). Jurassic World featuring Coca Cola and Avengers Endgame featuring Audi were chosen as they were popular films that contained product placement that both younger and older adults would recognise (Malaj, 2022), minimising the effects of familiarity. Furthermore, these two film clips were chosen as they contained product placement of products of different monetary value products. Thus, controlling for the potential effects of monetary value on product placement susceptibility (McDermott et al., 2006).
Both film clips were downloaded from Youtube and trimmed to last approximately one minute each to lessen the study length because of the propensity for individuals with PD to tire because of the symptomology they present with (see Appendix A for the screen shots of the two film clips). The two film clips were shown on a HP ProBook 470 G5 17.3” laptop because the large screen enhanced participants’ visual experience of product placement (HP, 2022).
Measure of Purchase Intention
Separate pre and post product questionnaires for each clip were made using Qualtrics (Qualtrics, 2022). To measure purchase behaviour, participants were asked how strong their preference was to buy those drink/car brands on a Likert scale of one to seven (from one = “Extremely unlikely” to seven = “Extremely likely”). Literature has found 7-point Likert scales to be a more reliable scale because it allows for more accurate and differentiated responses than smaller scales like 5-point Likert scales (Cicchetti et al., 1985; Finstad, 2010). The use of a 7-point Likert scale therefore gained a more sensitive and accurate measurement of product placement susceptibility. Both the pre and post product placement questionnaires asked participants the same questions therefore enabling us to measure if there was a change in participants’ responses prior to and after exposure to product placement (Matthes et al., 2007).
Design
The study used a 3 between (Participant Status: Healthy Young Controls vs. Healthy Older Controls vs. Individuals with Parkinson’s Disease) x 2 within (Product Placement Category: Drink vs. Car) mixed-subjects design.
Procedure
As this study recruited a vulnerable population, the information sheet was sent to participants via email 48 hours prior to the in-person study. This afforded participants the time to ask questions or express any concerns about the study before then being sent the consent form 24 hours prior to commencing the in-person study. Once participants had read and completed the digital consent form, participants were sent the digital HADQ. The HADQ took participants approximately 10 minutes.
Prior to the main study, participants were screened for cognitive impairment, using the ACE, and visual impairment, using the Ishihara test. At this time the severity of Parkinson’s symptomology was assessed using the MDS-UPDRS where appropriate.
On completion of all pre-study screening, participants were asked to firstly complete a prosaccade eye tracking task and then an antisaccade eye tracking task which took approximately 10 minutes.
Participants were then asked to complete a pre product placement questionnaire and then watch a short film clip. After watching the film clip, participants were asked to complete a post product placement questionnaire. Finally, participants were asked to complete the Stroop test which took approximately five minutes to provide a further measure of IC and to act as a buffer in time.
This process was repeated for a second product category condition. The order of condition completion was randomly counterbalanced across participants to increase internal validity by minimising the potential for order effects (Corriero, 2017). The in-person study lasted approximately an hour for healthy controls and an hour and 30 minutes for PD. At the end of the study, participants were read and given a copy of the debrief sheet, thanked for their participation and time, and given £10 as a contribution towards travel expenses. All raw data was stored on the Lancaster University OneDrive, on a password-protected computer.
Data Analysis
The raw data from the prosaccade and antisaccade tasks were extracted using the EyeLink DataViewer Software (Version 3.2) and processed using the bespoke software SaccadeMachine (Mardanbegi et al., 2019). Noise in the dataset was removed by filtering out frames with a velocity signal greater than 1,500 deg/s or with an acceleration signal greater than 100,000 deg2/s. The EyeLink Parser was used to detect fixations and saccadic events. Saccades were extracted alongside multiple temporal and spatial variables. Trials were excluded in cases when the participant did not direct their gaze to the central fixation target. The onset of target display was a temporal window of 80-700ms, thus anticipatory saccades made prior to 80ms and excessively delayed saccades made after 700ms were removed.
To improve data analysis reproducibility, statistical analyses were conducted using RStudio (version 2022.09.0) (Quick, 2010). To prepare the Stroop test data for analysis, participants’ Stroop scores (correctly identified ink colour out of 40), congruent and incongruent trial reaction times (ms), and Stroop effect (incongruent trials reaction time (ms) minus the congruent trials reaction time (ms)) were downloaded from Psytoolkit into an Excel file. IC was operationalised as the Stroop effect (Kane & Engle, 2003).
To investigate the susceptibility to product placement, a difference in purchasing behaviour score was calculated for each product. To do so, the pre product placement ratings of the likelihood of purchasing each brand were subtracted from the post product placement ratings of the likelihood of purchasing each brand. A positive difference was indicative of participants being more likely to buy the featured product after exposure to product placement, a negative difference suggested that participants were less likely to buy the featured product, and a difference of zero indicated no change in purchase behaviour.
First to confirm the assumption that is impaired in individuals with PD compared to healthy controls, three separate between-factor ANOVAs were performed to compare the main effect of group (YC, OC, and PD) on antisaccade latency, antisaccade error rate, and Stroop effect (See Appendix B for R code). A between-factor ANOVA was chosen because it compares three or more categorical groups to establish whether there is a significant difference on a dependent measure (Henson, 2015). As ANOVA results only identify a difference between groups, post hoc Tukey HSD tests for multiple comparisons were conducted to determine where the differences lie between groups (Abdi & Williams, 2010).
To investigate whether IC influences product placement susceptibility, a linear mixed effects modelling (LMM) was fitted. The LMM fitted incorporated difference in purchase behaviour scores (differencescore) as the outcome, and group (PD v Healthy older control v Healthy younger control) and measures of IC (antisaccade latency, antisaccade error rate, and Stroop) as the fixed effects. Given that IC is part of an individual’s executive function (Crawford et al., 2002), ACE score (as a measurement of the participants overall cognitive function; Noone, 2015) was also fitted as a fixed effect. As LMM allows for the analysis of fixed effects of independent variables, whilst also considering unexplained differences corresponding to random effects like participant variation (Baayen et al., 2008). Random effects of both participants and product (Car or Drink) on intercepts were added (See Appendix C for R code). The LMM was fitted using the Satterthwaite adjustment method in lme4 package (Bates et al., 2014) in R (version 2022.09.0) (Quick, 2010).
Ethics
This study received ethical approval from the Psychology Department Research at Lancaster University on the 22/06/2022 and complied to The British Psychological Society’s guidelines (2014).

Lancaster University

Data/R.csv

Ball2022

Elena Ball

Open

N/A

English

Data

LA1 4YF

Dr Megan Readman

MSc

Psychology of Advertising

53 Participants. 20 healthy younger controls, 20 healthy older controls, 13 individuals with mild-moderate Parkinson's disease

ANOVA
Linear Mixed Effects Modelling