MARK4040 Marketing Research: Data Preparation and Analysis

Data Preparation and Analysis

• The marketing research process involves 6 steps: problem definition, approach development, research design formulation, fieldwork/data collection, data preparation and analysis, and report preparation and presentation.
• Data preparation and analysis involves 4 major steps: data validation, editing and coding, data entry, and data tabulation.
• Data tabulation can be one-way (focusing on a single variable) or cross-tabulation (categorizing the number of responses for two or more variables).

Data Types

• A variable is an attribute whose value can vary.
• A constant is an attribute whose value does not change.
• A case (or observation) involves all observed variables associated with an object.
• A latent construct is a variable that is not directly observable.

Qualitative Data Analysis

• Major steps in qualitative data analysis: data reduction, data display, and conclusion drawing/verification.
• Data reduction involves coding and labeling the text to categorize it.
• Data display involves summarizing the data and showing the major ideas.
• Conclusion drawing/verification involves drawing conclusions and checking for possible biases.

Quantitative Data Analysis

• Measures of central tendency: mean, median, and mode.
• Mean is the average value within the distribution, suitable for interval or ratio data.
• Median is the middle value of the distribution, suitable for ordinal data.
• Mode is the value that appears most often in the distribution, suitable for nominal data.
• Measures of dispersion: range, standard deviation, and variance.
• Range measures the spread of values in a variable.
• Standard deviation refers to the average distance of the values of a distribution from its mean.
• Variance is the average squared deviation or the square of the standard deviation.

Hypothesis Testing

• Hypothesis testing is based on the idea of contradiction: assuming the opposite of what we want to prove and then trying to reject it.
• The null hypothesis (H0) states that there is no effect or no difference.
• The alternative hypothesis (H1) states that there is an effect or a difference.
• A p-value is obtained when running a statistical test, which is compared to a significance level (e.g., 5%).
• Parametric statistics (e.g., t-test, ANOVA) are used when variables are measured using an interval or ratio scale with a relatively large sample size.
• Nonparametric statistics (e.g., Chi-square test) are used when variables are measured using a nominal or ordinal scale.

Univariate and Bivariate Statistical Tests

• Univariate statistical tests (e.g., one-sample t-test) are used to test a hypothesis about a single variable.

• Bivariate statistical tests (e.g., Chi-square test) are used to test the relationship between two variables.### Paired Samples t-test

• Used to compare two means when the two variables (Q1 and Q2) are interval variables

• Null hypothesis (H0): no difference between two means

• Alternative hypothesis (H1): two means are significantly different

• In Excel, select Data > Data Analysis > t-Test: Paired Two Sample for Means

Running Paired Samples t-test in Excel

• Remove blank cells in both columns by filtering (Data > Filter)
• Copy the filtered data to a new sheet using Copy > Paste Special > Values (V)
• Select Data > Data Analysis > t-Test: Paired Two Sample for Means
• Specify the positions of the variables of interest in the Range boxes
• Set Hypothesized Mean Difference to 0
• Optional: select “Labels” if question label is in the first row
• Click OK

Interpretation of Paired Samples t-test

• t-value and p-value are obtained
• If p-value < 0.001, reject the null hypothesis that two means are equal at the 1% significance level

One-way ANOVA

• Used to compare group means
• Null hypothesis (H0): no difference between group means
• Alternative hypothesis (H1): group means are significantly different
• In Excel, select Data > Data Analysis > Anova: Single Factor
• Specify the positions of the variables of interest in the Range boxes
• Optional: select “Labels in First Row” if question label is in the first row
• Click OK

Running One-way ANOVA in Excel

• Exclude all empty/blank cells in the dependent variable Q4
• Create new columns containing values of Q4 for each specific group in Q1
• Repeat Steps 3 to 5 for each group
• New data should look as follows:

Interpretation of One-way ANOVA

• F-value and p-value are obtained
• If p-value > 0.05, cannot reject the null hypothesis that group means are equal at the 5% significance level

Correlation Matrix in Excel

• Used to obtain the correlation matrix between variables of interest
• Exclude all empty/blank cells from all variables of interest
• Select variables of interest, then click Data > Data Analysis > Correlation
• Verify the positions of the variables of interest in the Range boxes
• Optional: select “Labels in First Row” if question label is in the first row
• Click OK

Interpretation of Correlation Matrix

• Correlation coefficient between Q4 and Q5 is obtained
• Correlation coefficient is always between -1 and 1
• Negative correlation indicates a negative linear relationship between two variables
• Positive correlation indicates a positive linear relationship between two variables