Text Analysis

A company is using text analytics to study Google reviews of their products. The goal is to identify customer suggestions for product improvements. The company followed a methodology (based on Zingle et al., 2019) to label each review with three variables:

• The product mentioned in the review
• The presence of recommendations (whether the review includes a suggestion for improvement)
• The sentiment of the review (positive or negative, ranging from -1 to 1).

The company used the following steps:

• They identified patterns in the text, particularly looking for modal verbs followed by other verbs, as indicators of recommendations.
• They conducted sentiment analysis on each review, assigning a value between -1 and 1 to represent the overall sentiment of the review.
• They classified reviews as either requiring action (negative reviews with suggestions) or not requiring action (positive reviews with suggestions).
• They focused on the most mentioned products in the reviews that contained negative sentiment with suggestions for improvement.

Step-by-Step Analysis:

Your task is to analyze each step of the methodology and:

• Identify whether deductive or inductive logic was used at each step.
• Point out any logical flaws, incorrect assumptions, or methodological biases in each step of the process.

Specifically:

• Was the decision to use modal verbs followed by verbs as a rule for identifying recommendations based on deductive or inductive reasoning? Do you think this rule is reliable?
• Does the use of sentiment analysis provide an accurate assessment of customer satisfaction? Could it be biased? Does sentiment analysis work effectively with reviews that may contain sarcasm or indirect language?
• Was the decision to focus only on negative reviews with suggestions logically sound? How might this approach introduce bias in identifying which products need improvement?
• Is the methodology biased in any way? Consider how the type of feedback (positive vs. negative) and focus on certain products might affect the results.

Business Case & Bias Connection:

• Identify any biases in the methodology, especially those related to data analysis and decision-making. How might these biases affect the conclusions drawn from the reviews?
• Connect the bias in the methodology to the broader concept of logic. Does the reasoning used lead to potentially skewed or incorrect conclusions? Explain.

Report Writing:

• Write a 1-2 page report analyzing the methodology used in the case study.
• Discuss the use of deductive and inductive logic in each step.
• Identify flaws, assumptions, or biases in the methodology and explain their potential impact on the results.