Market Research- Topic 8

What is the purpose of questionnaire checking?

To identify incomplete, unacceptable questionnaires.

What is the purpose of editing?

To detect illegible, inconsistent, or ambiguous responses.

What is precoding?

Assigning codes in advance for structured questions.

What is postcoding?

Assigning numerical codes to open‑ended responses.

What is transcribing?

Transferring coded data into a computer or software.

What is data cleaning?

Detecting and correcting errors, handling missing data.

Why might data need statistical adjustment?

To make the sample more representative of the population.

What is the final step of data preparation?

Selecting a data analysis strategy.

How is a one‑answer question entered in Data entry procedures (SPSS)?

1 question = 1 variable.

How is a multi‑response question entered in SPSS?

1 response alternative = 1 variable.

What elements must be defined for each SPSS variable?

Name, label, values, value labels.

What should data analysis be based on?

Research questions, data characteristics, statistical technique properties, researcher background.

What are the two main types of data?

Non‑metric (nominal, ordinal) and metric (interval, ratio).

What are univariate techniques used for?

Analysing one variable.

What univariate technique is used for non‑metric data?

Frequency distribution.

What univariate technique is used for metric data?

Mean.

What are multivariate techniques used for?

Analysing two or more variables.

BIVARIATE TECHNIQUES

Aim: Influence

What technique is used when at least one variable is non‑metric?

Cross‑tabulation (chi‑square).

What technique is used when both variables are metric?

Correlation (Pearson coefficient).

BIVARIATE TECHNIQUES

Aim: Comparison

What technique is used to compare two subsamples?

t‑Test.

What technique is used to compare two or more subsamples?

ANOVA.

What is the null hypothesis (Ho)?

A statement of no effect, no difference, or no association.

What is the significance level typically used?

0.05.

What does Pr < 0.05 mean?

Reject Ho.

What does Pr > 0.05 mean?

Do not reject Ho

Condition for chi‑square?

At least one variable is non‑metric.

Purpose of chi‑square?

To test association between variables.

Null hypothesis for chi‑square?

Ho: There is no association between variables.

Decision rule for chi‑square?

• Pr < 0.05 → reject Ho → variables are associated

• Pr > 0.05 → do not reject Ho → variables are independent

Condition for correlation?

Both variables must be metric.

Purpose of correlation?

To measure association between variables.

Null hypothesis for correlation?

Ho: Correlation coefficient = 0 (no association).

Decision rule for correlation?

• Pr < 0.05 → reject Ho → variables are associated

• Pr > 0.05 → do not reject Ho → no association

When is an independent samples t‑test used?

When comparing means of two different groups.

Null hypothesis for t‑test?

Ho: Means are equal.

Decision rule for t‑test?

• Pr < 0.05 → reject Ho → means differ

• Pr > 0.05 → do not reject Ho → means are equal

When is a paired samples t‑test used?

When the same respondents are measured at two different times.

Null hypothesis for paired t‑test?

Ho: Means are equal.

Condition for ANOVA?

Dependent variable = metric; independent variable = non‑metric.

Purpose of ANOVA?

To compare means across 2+ groups.

Null hypothesis for ANOVA?

Ho: All group means are equal.

Decision rule for ANOVA?

• Pr < 0.05 → reject Ho → means differ

• Pr > 0.05 → do not reject Ho → means are equal