Causal Research Methods

• Research design encompasses various methods for conducting research
• Includes causal, correlational, experimental, market analysis, financial analysis, and sampling techniques

Causal Research

• Involves instances where one variable directly or indirectly influences another
• Includes the tracing of an unbroken chain of influence from initial event to final result
• Causal relationships can be unidirectional (A influences B but not vice versa) or bidirectional (each variable influences the other)

Strengths

• Identifies causes behind processes, which aids in problem-solving and outcome optimization
• Replication is possible
• Helps to identify the impacts of changing the processes & exiting methods
• Involves systematic subject selection, enhancing internal validity

Limitations

• Difficult to administer due to challenges in controlling extraneous variables
• Can be expensive, sometimes costing over $1–2 million for real-life advertising campaign tests
• Reveals plans to competitors, who might use research outcomes to preempt market entry
• The sole outcome of causal research is insufficient; requires additional research for output confirmation

Correlational Research

• Primarily focuses on determining if variables covary, establishing directions, magnitudes, and forms of relationships
• Focuses on observing variables "as is" without manipulation
• Some approaches compare a variable's average value across preformed groups

Strengths

• Enables more data collection compared to experiments
• Results tend to have greater real-world applicability since research occurs outside the lab in normal conditions
• Provides a starting point for new research and allows the determination of a relationship's strength and direction

Limitations

• Only uncovers relationships; cannot determine why a relationship exists
• It does not reveal which variable influences the other
• Cannot assert causation definitively without additional research

Experimental Research

• Includes manipulating independent variables, and controlling extraneous variables

Manipulation of Independent Variables

• An independent variable's values are chosen by the experimenter
• Two levels of a variable should be present to manipulate an independent variable
• Effects of manipulation on behavior are then recorded
• The dependent variable's value depends on the level of the independent variable and is observed/measured

Control Over Extraneous Variables

• Extraneous variables may affect the behavior of interest but aren't part of the study
• Uncontrolled extraneous variables can cause uncontrolled changes in the dependent variable, obscuring the independent variable's impact
• Controlling these variables involves holding them constant, or randomizing their effects across treatments

Strengths

• Identifies and describes causal relationships

Limitations

• Restriction of use: The experimental method cannot be used for hypothesized causal variables.
• It would unethical to expose people to various conditions to identify which of those conditions cause personality disorders.
• Tight control over extraneous factors reduces the ability to apply findings to real world situations.

Market Analysis

• Helps determine action course for implementing a strategic plan
• Includes quantitative and qualitative data for a picture of tangible and intangible factors

The Four Parts

• Industry overview, such as size, trends, and projected growth
• Target market demographics, location, psychographics, behaviors, and trends
• Competition, is broken down into direct and indirect, includes competitive advantage and barriers to entry
• Pricing and forecast helps determine a business's position in the market

Financial Analysis

• Evaluates businesses, projects, and budgets to check performance and suitability while assessing company solvency
• Used to evaluate trends, set policy, make long-term plans, and identify investment projects
• Done by synthesizing numbers/data and conducting corporate/investment finance activities
• Financial statements and ratios are thoroughly examined for their historical performance

Basic Financial Ratios

• Working capital ratio assesses if a business can pay current debts and obligations
• Quick ratio assesses how current liabilities are covered by assets
• Earnings per share (EPS) measure the net income earned on each share
• Price-earnings ratio (P/E) reflects assessment of future earnings based on stock price and EPS

Competitor Analysis

• Place competitors in strategic groups to see how directly they compete
• Also list their product/service, profitability, growth pattern, objectives, strategies, structure, strengths, weaknesses, and size

Sampling Techniques

• Used to gather information for decision-making
• Data is collected from samples to make conclusions about a larger population

Reasons for Sampling

• Sampling saves on time and money
• Given resources, sampling broadens study scope
• Sampling preserves products
• If access to the population is impossible, sampling is the only suitable option

Random Sampling Methods

• Simple random sampling: a draw from the sampling frame where each member has equal selection probability
• Stratified sampling: population divided into homogeneous groups, then samples taken from each
• Cluster sampling: population divided into non-overlapping areas, good representatives for the entire region
• Systematic Sampling: a basic technique used because of its ease of adminstration
• Multistage sampling: selection occurs in two+ steps; helpful when no sampling frame covers the entire population of interest

Non-Random Sampling Methods

• Quota sampling: subclasses such as age, gender, and location, are strata with non-random sampling to gather data until subclass quota is filled
• Convenience sampling: Select elements for convenience to the researcher
• Judgment sampling: select elements that researcher deems best fit for study
• Snowball sampling: Uses referrals from other survey respoondents because they are hard to locate

Sampling Distribution; Central Limit Theorem

• (1) as sample size increases, the shape of the sampling distribution of means approximates normal,(2) the mean of the sampling distribution of means will equal the population mean, (3) the standard deviation will equal √(σ/√n)
• Shape: Even if the population is not normal, make the sample size larger, like 60+, to make it normal.
• • The standard error is called "error" because any particular sample mean is expected to be close to the population mean, but is not expected to be exact
• 𝜎𝑀 is Expected to vary from μ

Questionnaire Design

• Guidelines for questionnaires ensure responses from potential respondents
• Questions reflect the research objective and should be adapted and unbiased based on culture

Principles of Wording

• The appropriateness of the questions' content: Tap into subjective variables such as attitudes with scaled ordinal sets
• How questions are worded and the language level used: Should approximate the respondent's knowledge, usage of terms, and frames of reference
• • Open-ended questions allow respondents to answer in the way they choose unlike close ended questions that ask respondents to make quick decisions by choosing mutual exclusive categories

Question Types

• Use both + and - worded questions to minimize response bias
• Double-barreled questions are best avoided; separate them into two or more separate questions
• Avoid Ambiguous questions
• Questions must also not be dependent on recall
• Leading questions should be avoided to not lead the respondents to agree or disagree with the questions intent

Other Guidelines

• loaded questions are bad
• word questions so that they do not elicit high social desirability scores
• keep it simple, and keep questions short
• place question with the easy questions first, and progressively get more difficult
• do not put like questions directly after each other because it could be insulting

Principles of Measurement

• The scales/scaling techniques used in question measuring are connected to minimizing bias
• To ensure appropriate data data is collected to test the hypothesis, use correct scales depending on data
• Interval and ratio scales should be used on preference to nominal and ordinal
• Check tests for validity and reliability
• Make sure it's a "general get up" to also pay attention to how the questionnaire looks with neat intro, instructions and questions