Quantitative Data Collection: Observations

Observational research in quantitative study designs share similarities with qualitative ethnographic studies, where researchers observe individuals in their natural settings. It involves researchers observing participants' behaviours in natural or normal environments to gain valuable insights. However, quantitative observational research focus on measurable data, such as the number of people performing a particular action, the type of action being performed or the quantity of content or a product is picked up. These studies are typically structured, requiring observers to document behaviours according to predefined criteria. Structured observation demands a focus on specific and often subtle behaviours to uncover authentic patterns and behaviours. Observers must exercise their judgement to classify actions, such as reading product labels before making a selection, comparing items before deciding, or choosing based on price considerations. This technique can be conducted in two ways: Human Observation or Automated Observation (Davis, 2011).

Human Observation

In human observation, researchers monitor participants' behaviours, either by immersing themselves in the environment or by observing remotely or in controlled laboratory settings. The researcher may either actively engage with participants or remain non-intrusive. Therefore, human observation is more subjective than other quantitative data collection tools, yet in some situations, it may be the only way to collect the necessary data (Sudan, 2017). Human observation is particularly effective in four scenarios (Davis, 2011):

When observing behaviours provides deeper insights than verbal descriptions; when participants struggle to articulate attitudes; when survey methods fail to accurately represent human behaviours; and when behaviours themselves are the primary source of insight.

Human observation is widely used in areas like scientific observation, physiological research, and market or product studies. Researchers interpret observed behaviours as either quantitative data (numerical) or qualitative data (patterns or categories).

Human observation relies on four key aspects (Davis, 2011):

1. Situation: Natural (real-world settings) or vs. Artificial (controlled environments).

2. Observer Obtrusiveness: Open (participants know they’re being observed) vs. Disguised (participants are unaware).

3. Observer Participation: Active (direct interaction with participants) vs. Passive (no interaction).

4. Data Recording: Structured (predefined criteria) vs. Unstructured (open-ended observations).

Natural vs. Artificial Settings

Natural Settings: In the social sciences, natural observational research refers to a study design conducted outside the laboratory to observe the behaviour, particularly in participants' real-world environments where they behave naturally (Crano et al., 2011). For example, observing customers in a restaurant to study behaviours based on gender or race; monitoring sales clerks' interactions with customers in various retail outlets; or examining media consumption habits in a doctor’s waiting room. There are three key components necessary to natural observation: the natural setting, natural event, and natural behaviours.

This method offers significant strengths: It helps interpret complex or novel situations; it allows researchers to measure authentic behaviours; it enables the development of new theories. The limitations of natural observation include researchers having no control over the environment and it is unsuitable for certain types of observation. It can also be challenging to record or observe all occurrences in a natural setting.

Artificial Settings: Artificial observation occurs in controlled environments, such as laboratories, where specific scenarios are created to study behaviours. An example is observing participants’ use of traditional or social media in a specific room or laboratory setting. Artificial observation involves creating fabricated situations and manipulating content to testing consumer responses to these scenarios.

Compared to natural observation, artificial settings offer two key advantages. It speeds up the data recording and gathering process and allows researchers to control external factors, enhancing consistency.

Open vs. Disguised Observation

In open observation, the researcher’s role is transparent and participants are aware their activities are being observed, followed and recorded (Davis, 2011). In disguised observation, the researcher’s presence hidden, ensuring participants remain unaware of the observation (Gardner, 2000; Boote & Mathews, 1999). Disguised observation is especially useful in market research as it often yields more realistic behavioural data than open observation.

Active vs. Passive Observation

Active observation involves the researcher directly engaging with participants, asking questions, or interacting as part of the observation. In contrast, passive observation entails silently monitoring participants without interaction, similar to a disguised observer (Cohen et al., 2017).

Each approach has its pros and cons. Active observation can collect detailed data through direct interaction but risk introducing bias. Passive observers avoid bias but cannot interact, which may limit the depth of data collected. Passive observation is time-intensive but often results in more authentic behavioural data.

Structured vs. Unstructured Observation

Structured Observation: Structured observation relies on predefined criteria, such as a checklist, to guide data collection during observation. The researcher should carefully decide what to observe, how to observe, how long, and how to record the observed data (Sadan, 2017). It is typically used in quantitative research. While preparation and data collection require considerable time and effort, structured data is easier to analyse after data collection as everything is structured.

Unstructured Observation: Unstructured observations are done spontaneously and recorded as what is seen (Sadan, 2017). This allows for greater flexibility, enabling researchers to record observations in narrative form. It is particularly useful for uncovering new knowledge but requires more time for data coding and analysis.

Automated Observation

Automated observation uses technology to monitor participants’ behaviours instead of relying on human observers (Davis, 2011). It can be conducted in three methods: Direct monitoring (online/offline), observing consumer-generated content, and eye tracking.

• Direct Monitoring: Observing online or offline behaviours through digital tools.

  
  

• Consumer-Generated Content: Analysing user data, such as interactions on websites or social media.

  
  

• Eye Tracking: Recording eye movement and gaze patterns to understand attention and interest (Carter & Luke, 2020).

Websites are able to offer information about the preferences of its users by monitoring, storing, and analysing the interaction data of its users. Facebook, Google, and YouTube leverage automated observation to monitor and analyse user data, optimise content delivery based on their interactions, and tailor advertisements to individual preferences (Zdziebko & Sulikowski, 2015). These are done directly in the observing consumer-generated content process.

Eye Tracking

Eye tracking is a specialised technique where devices measure eye motion and gaze position over time. This method is widely used in both online and offline advertising to identify customer interests and adjust content accordingly. Recent advancements in eye-tracking technology have improved the accuracy of these observations, making it a valuable tool for understanding consumer preferences.

Commonly used methods in human observations are check lists and rating scales (Polit & Beck, 2017):

Category Systems and Checklists

In observational research, behaviours, events, or attributes of participants are grouped into predefined categories for observation and recording. These categories must be clearly defined to ensure accuracy and consistency. For effective observation, it is recommended to limit the number of categories to 15–20 (Grove et al., 2013). Observers then interpret the recorded data based on these categories.

Participant behaviours are monitored to determine whether specific actions occur, with occurrences recorded as tally marks within the relevant categories. In a checklist, a single category is typically selected for observation, and the checklist is created using a category system. For example, when assessing behavioural indicators of pain, categories might include crying and facial expressions. Facial expressions are further broken down into specific indicators, such as brow bulging, eye squeezing, or the presence of a nasolabial furrow, with each checked off as observed.

Rating Scales

Rating scales allow observers to evaluate a participant's behaviour or an event on a numerical or descriptive scale at set time intervals, enabling the data to be quantified. Compared to checklists, rating scales provide more detailed information. When combined with a category system and checklists, the data collected becomes significantly more valuable for analysing the phenomenon being studied. Rating scales can be utilised for both observational purposes and self-reporting by participants.

Conclusion

Observational research is a versatile method that bridges quantitative and qualitative approaches. Whether conducted through human observation or automated systems, it provides valuable insights into participant behaviours in natural and artificial settings. While each method has its strengths and limitations, combining structured techniques with modern technologies like eye tracking enhances the depth and accuracy of findings. By selecting the appropriate observational approach, researchers can unlock authentic insights that inform better decisions across fields like market research, social sciences, and beyond.