Research Approach

research sample consistent with the delimitation used in previous studies facilitates a more precise interpretation of the results. The choice for gender-specific delimitation is discussed in further depth in the Limitations section in Chapter 5.6. and also addressed in the Future Research section 5.7.

Compulsive Buying Assessment. Based on the psychometric properties of the scales, Compulsive Buying Scale from Valence et al. (1988) is chosen as the primary method for assessing the consumer’s compulsive buying tendency. The Compulsive Buying Index (Ridgway et al., 2008) is additionally employed to validate the chosen instrument’s convergent validity.

Based on indicated scores, the consumers are divided into two groups: those with a high compulsive buying tendency and those with a low compulsive buying tendency, which are hereafter respectively also referred to as “compulsive buyers” and “non-compulsive buyers”/

“prudent buyers”. The specific methodological choices employed in each study are further presented in Chapter 3 (Study I) and Chapter 4 (Study II). The limitations of these choices are outlined in Chapter 5.6.

To test the deducted hypotheses, neurophysiological and physiological data in combination with verbal data are collected. Stationary and mobile eye-tracking devices are used to measure physiological responses, namely, visual attention, to the stimuli that are tracked through the visual system. An electroencephalograph (EEG) is employed to study electromagnetic brain activity, thereby enabling the investigation of cognitive responses such as engagement and cognitive workload. A biosensor measuring electrodermal activity (EDA) is employed to track the changes in physiological arousal. Self-reports with both open and closed questions are used to better understand the consumer’s opinions of the constructs under investigation.

Eye-tracking. Eye-tracking is a valuable technique to identify and measure the immediate physiological responses induced by the presented visual information. Eye movement helps us determine the sequence of data selection and acquisition, and it also provides information on the temporal aspects of the studied cognitive processes (Duchowski, 2002, 2007; Gidlöf et al., 2013;

Holmqvist et al., 2011). We can track the location of a subject’s gaze using eye-tracking devices, which use corneal reflections induced by infrared light to locate the positions of the pupil and cornea, from which we can estimate the point of a gaze in a presented image (for technical details, see: Duchowski, 2007; Holmqvist et al., 2011). Although eye-tracking can

10

give important temporal and spatial indication of the cognitive processes, it does not provide any information about which cognitive processes are involved (Holmqvist et al., 2011).

Electroencephalography (EEG). An EEG is a neurophysiological measurement that can capture complex patterns in brain activity in milliseconds after being exposed to a stimulus (Teplan, 2002; Zurawicki, 2010). Rather than measuring the deep brain structures, EEGs record brain activity in frequencies by capturing the electric field in the scalp. The amplitudes of the electric currents are related to frequency bands of mental states. Although an EEG can provide a precise understanding of the slightest neurophysiological changes induced by stimuli and feature a high temporal resolution, it has a low spatial resolution (Venkatraman et al., 2015; Zurawicki, 2010).

Hence, it is often coupled with other physiological measurement tools.

Galvanic skin response (GSR). This technique is also known as skin conductance (SC) or electrodermal activity (EDA). GSR captures changes in sweat secretion by non-invasively recording the skin’s electrical characteristics and providing the measures in micro-Siemens (μS) per unit time (Boucsein et al., 2012; Boucsein, 2012; Fowles et al., 1981). The latency of the response is slow since reactions are recorded 1-2 seconds after the onset of the stimulus. GSR responses are important to consider because sudomotor activity plays a dominant role in thermoregulation and sensory discrimination (Boucsein, 2012). Specifically, the secretion of sweat in hands and feet is a robust indication of emotional stimulation representing activation of the autonomic nervous system, which could indicate an orienting response or a more general emotional arousal (Boucsein, 2012; Ravaja, 2004). EDA responses are hard to be consciously controlled since they are modulated autonomously by sympathetic activity which motivates human behavior through cognitive and emotional states under conscious awareness (Boucsein, 2012). Measuring of tonic and phasic EDA responses enable us to assess the changes of emotional and motivational components of arousal that reflect the relevance of a stimulus or an event in the surrounding environment (Boucsein, 2012). However, although the EDA parameters can indicate slightest variations in arousal, they cannot provide information about the valence or subjective interpretation of the experienced emotional state. To measure the simultaneous perception of the stimuli in the natural environment (e.g., store), by capturing the variations of physiological changes in emotional responses, a mobile EDA device is often employed (Bagozzi, Gopinath, & Nyer, 1999; Groeppel-Klein, 2005; Groeppel-Klein & Baun, 2001). This data-collection method has been validated and successfully applied in other affect

11

investigating studies in the consumer research domain ( Groeppel-Klein, 2005; Groeppel-Klein

& Baun, 2001; Ohme, Reykowska, Wiener, & Choromanska, 2009).

By coupling the eye-tracker with other tools to collect neurophysiological and physiological data, such as EEG or EDA measurement devices, the researcher can track the changes in the measurements during the time frame of interest and at the specific position. A more detailed description of the specific methods employed for each study is provided in Chapter 3 and Chapter 4.

The integration of state-of-the-art measurement techniques (i.e., EEG, eye-tracking, and EDA) in this study enables the recording of pertinent behavioral signals in the investigated scenarios. It also complements and advances traditional approaches because consumer neuroscience tools that measure biofeedback can help overcome the limitations of data-collection techniques based on self-reporting (Kenning, Plassmann, & Ahlert, 2007; Knutson, Rick, Wimmer, Prelec, &

Loewenstein, 2007; Plassmann et al., 2012). Compulsive buying is a sensitive topic. Self-report measures can often be influenced by various biases as consumers may often provide strategically-shaped responses affected by filters of sense and/or social desirability, they may be reluctant, or unable to verbalize their behavior or experienced states (Dimofte, 2010; Nevid, 2010). By integrating the consumer neuroscience methods to collect data, this study offers a potential to explain the variance in the studied phenomenon at a more in-depth level, which is necessary for the development of more neuropsychologically sound theoretical models (Kenning et al., 2007; Knutson et al., 2007; Plassmann et al., 2012; Solnais, Andreu-Perez, Sánchez-Fernández, & Andréu-Abela, 2013; Yoon et al., 2012)

In addition, by conducting a field experiment (Study II), this thesis responds to the emerging call for studies that use mobile data-collection methods to investigate consumer behavior and choices in more natural environments (e.g., Gidlöf et al., 2013). This offers pioneering theoretical and methodological insights that could be valuable for further research attempts, both in academic and commercial fields.

12

In Study I, the cognitive responses to brand advertisements in two groups of consumers with a high and low compulsive buying tendency are examined during a task that simulates the viewing of a TV-advertisement. The advertising variables underlying information processing (namely, engagement and cognitive workload during ad-exposure, visual attention to visual brand elements, and self-reported memory performance) are estimated and compared between two groups. Additionally, the impact of the presented advertising category and the relationship between the attention measures and the memory measures for the two groups are also explored.

In Study II, emotional and behavioral responses in two groups of consumers — again grouped into a high or low compulsive buying tendency groups— are tested during a shopping excursion through two single-brand fashion-apparel stores, one representing a low-end fashion store and one a high-end fashion store. The variables that indicate the effects of the store during interaction with the in-store environment (namely, arousal and behavioral shopping-experience outcomes) are measured during a shopping experience in each fashion store and compared between the two groups. The relationship between the emotional and behavioral measures and their differences depending on the groups are also explored.

An overview of the research studies included in the thesis is presented in Table 1-1 below.

Setting Constructs Consumer segments Research methods Research design Analyses

Study I Pre-purchase:

Interaction with brand advertising

Attention and memory

Consumers with a high and low compulsive buying tendency

EEG

synchronized with stationary

eye-tracker and self-report- based

questionnaire

Simulated laboratory experiment

Linear Mixed Models Generalized Linear Mixed Models Parametric and nonparametric tests Spotlight analyses

Study II Purchase:

Interaction with the store environment

Arousal and in-store behavior

Consumers with a high and low compulsive buying tendency

Mobile EDA tracking device, eye-tracking glasses and self-report- based

questionnaire

Field experiment

Linear Mixed Models Parametric and nonparametric tests Spotlight analyses

Table 1-1. Overview of studies included in this Ph.D. thesis

13