The Cochran-Armitage test of trend analysis

Get Complete Project Material File(s) Now! »

Choice of Method

The choice of method is a quantitative deductive research approach. It is used to answer the research questions regarding the relationship between lack of control over the situation, and negative emotions as well as testing the relationship between emotions and NWOM. Quantitative research is a part of the positivist paradigm because it adopts a highly precise, logical mathematical approach to collecting and observing numerical data (Collis & Hussey, 2014). A quantitative method analyzes the data through various statistical tests. The methods are well structured and often times aim to draw conclusions about hypotheses (Saunders, Lewis, & Thornhill, 2009). The research question at hand is seen as quantitative because it does not wish to draw conclusions about why negative emotions change when the customer’s control is limited, nor will the study describe the phenomenon from a subjective point of view. Those two types of studies are seen as an interpretivism approach, because of the methods required to conduct the studies (Saunders et al., 2009). The major setback to doing quantitative is that the research will not be able to answer why someone feels the way they do which is what interpretivism is able to do (Collis & Hussey, 2014). This lack of deeper interpretation may be seen as a negative factor when working with psychological customer behavior such as emotions.

Content of the survey

In the survey, five scenarios are designed to investigate the variables (lack of control, emotions, and NWOM) in more depth. In order to create the five scenario questions, a set of criteria has to be made during the creation of the survey. In this case, the set of criteria is a set of guidelines of what a service failure should entail in order to be considered one. Lind (2007) has proposed a list of criteria for an ethics study, but no such criteria has been made for service failures. The first criteria is that the scenario has to be a common occurrence within the airline industry and the second one is that it has to be a non-catastrophic failure. The third criteria is that the scenario must be clearly defined in one sentence in order to maintain organization during the analysis process. Finally, the scenario should have the possibility of eliciting some kind of negative emotion from the customer. Based off of those four criteans, five scenarios are made. The survey emphasizes on several service failure scenarios that are more common including: luggage handling service failure, delayed/cancelled flight, missed flight due to factors beyond passenger’s control, negative customer service at the airport, and negative service experience during the flight.

Gathering data

The first step of starting the process of conducting a survey was to create a trial survey with a few questions regarding five airline service failure scenarios that at the time were named as lost luggage, delayed/cancelled flight, missed flight due to factors beyond customer’s control, negative customer service at the airport, and negative inflight experience. The names were later on changed for the main survey in order to be more clear and easily understandable for nonnative speakers. The trial survey was first sent to family and friends where they needed to choose which scenario they have already encountered before, and then list the emotions that they have felt after either of those incidents. This was done in order to discover which emotions were the most experienced ones in regards to airline service failure, and then apply these emotions to the final survey. The most common emotions that were taken from the trial survey were anger, frustration, helplessness, nervousness, panic, and worry. With this information, the final survey was designed.

Airport Scenario

All negative emotions except for helplessness are rejected due to having no statistical significance difference (see Appendix 1, Table 4.5). This means that the analysis shows that the null hypotheses for anger, frustration, nervousness and worry can neither be rejected nor accepted. This scenario also finds a weak positive correlation between helplessness and feelings of lack of control over the situation, rs(98)=.318, p<.05. Panic is rejected in all the scenarios while helplessness is the only one that is accepted for further interpretation across all of the scenarios based on the significance levels. Within all the scenarios the data shows that null hypotheses for every emotion that went on for further interpretation would fail to reject the null hypotheses H1-H5 because they all have levels of positive correlation between emotion and lack of control over the situation. All of these results have positive correlations, but the Somers’ d is also conducted for cross referencing purposes to look for any discrepancies between the two tests.

Spearman Correlation

Spearman rank-order correlation, which is often just called Spearman correlation, is a nonparametric test that measures the strength of the association between two ordinal variables (Spearman, 2010). It can also measure whether the direction of the two variables is positive or negative (Chen & Popovich, 2011). There are three assumptions that need to be taken into consideration before doing a Spearman correlation. The first assumption is that the two variables 39 have to either be continuous or ordinal type variables. Ordinal is a type of variable that has two or more categories with clear ordering or rank. The second assumption is that there has to be a monotonic relationship. A monotonic relationship is one in which if variable A increases then variable B will increase as well.

Table of Contents :

  • 1. Introduction
    • 1.1. Background
    • 1.2. Problem
    • 1.3. Purpose
    • 1.4. Research Questions
    • 1.5. Perspective
    • 1.6. Delimitation
  • 2. Theoretical Background
    • 2.1. Causal Attributions
      • 2.1.1. Causal attribution theory
      • 2.1.2. Controllability
    • 2.2. Emotions
    • 2.3. Negative Word-of-Mouth
      • 2.3.1. Individual factors influencing NWOM
      • 2.3.2. Negative emotions and NWOM
    • 2.4. Theoretical framework
  • 3. Methodology
    • 3.1. Choice of Method
    • 3.2. Searching parameters
    • 3.3. Reasoning of choice to conduct a survey
    • 3.4. Content of the survey
    • 3.5. Gathering data
    • 3.6. Sample derivation and size
    • 3.7. Limitations
  • 4. Data Analysis
    • 4.1. General profile of the survey
    • 4.2. Factor Analysis and Reliability Test
      • 4.1.1. Factor Analysis and Reliability test results
      • 4.1.2. Interpretation of the Factor Analysis and Reliability test
    • 4.2. Descriptives of emotions
      • 4.2.1. Results of descriptive data
      • 4.2.2. Interpretation of the descriptive data
    • 4.3. Spearman Correlation
      • 4.3.1. Interpretation of results of Spearman Correlation
    • 4.4. Somers’ delta
      • 4.4.1. Interpretation of results of Spearman’s correlation and Somers’ d
      • 4.4.2. Hypothesis H1, H2 and H4 approval:
    • 4.5. The Cochran-Armitage test of trend analysis
      • 4.5.1. The findings for the luggage handling service scenario:
      • 4.5.2. The findings for the delayed/cancelled flight scenario:
      • 4.5.3. The findings for negative customer service at the airport:
      • 4.5.4 Interpretation of the Cochran-Armitage test of trend
  • 5. Conclusion
  • 6. Discussion
    • References

Airline Service Failures

Related Posts