Chapter 5: Organization of Data for Analysis

Numerical (Quantitative) data

Data in the form of any number.

Categorical (Qualitative) Data

Data that can be sorted into distinct groups or categories.

Ordinal Data

Qualitative data that can be ranked. Examples: Poor, fair, good, very good.

Nominal Data

Qualitative data that cannot be ranked. Examples: Blue Eyes, Green Eyes, Brown eyes

Primary Source Data

Data that have been collected directly by the researcher and have not been manipulated or summarized.

Microdata

An individual set of data about a single respondent.

Secondary Source Data

Data used by someone other than those who actually collected them.

Aggregate Data

Data that are combined or summarized in such a way that the individual microdata can no longer be determined.

Response Bias

When respondents change their answers to influence the results, to avoid embarrassment, or to give the answer they think the questioner wants.

Sampling Bias

When the sample does not closely represent the population.

Measurement Bias

When the collection method is such that the characteristics are consistently over or under-represented.

Non-Response Bias

When the opinions of respondents differ in meaningful ways from those of non-respondents.

Simple Random Sampling

A sampling method where every individual in the population has an equal chance of being selected, and choosing one person does not affect the chances of selecting another.

Systematic random sampling

A method where you choose a random starting point, then select individuals at regular intervals (e.g., every 10th person) from an ordered list.

Stratified Random Sampling

The population is divided into homogeneous groups called strata (such as age, gender, or grade level). A proportional simple random sample is taken from each group to ensure representation.

Cluster Random Sampling

The population is divided into heterogeneous groups (clusters) that each resemble the whole population. A random selection of entire clusters is chosen, and all members of those clusters are surveyed.

Multi‑stage Random Sampling

A complex method that uses multiple levels of random sampling. First, groups are randomly selected, then smaller groups within them, and finally individuals within those groups.

Destructive Sampling

A method where selected samples are destroyed or permanently altered during testing. It is used when testing requires breaking, consuming, or damaging the sample.

Multi‑phase Sampling

Sampling done in two or more steps. A simple, broad question is asked first to narrow the population, followed by more detailed questions for those who qualify.

Voluntary Sampling

Participants choose to take part on their own, usually by responding to an open invitation. Results often reflect strong opinions rather than the general population.

Convenience Sampling

The researcher selects individuals who are easy to access, such as people nearby or readily available.

Judgement (Purposive) Sampling

The researcher intentionally selects individuals who are believed to be most knowledgeable or best suited to provide the needed information.

Quota Sampling

Sampling continues until the researcher has collected a specific number of individuals from certain subgroups (e.g., 10 males and 10 females), without using random selection.

observational studies

A type of study where researchers observe situations that are already occurring and make inferences without manipulating any variables.

experimental studies

A study where researchers control variables and apply treatments to measure their effects.

treatment group

Participants in an experiment who receive the specific treatment being measured

control group

The group of participants that does not receive the treatment. It serves as a baseline for comparison with the treatment group.

Control (in experiments)

ensuring that all other factors in an experiment are kept constant so researchers can identify what caused any observed effect.

Randomization

The process of randomly assigning subjects to groups to prevent bias and ensure fairness in the experiment.

replication

Repeating an experiment with similar groups to confirm results and detect changes or patterns more reliably.

Sampling bias

Happens when the sample does not accurately represent the population. Certain groups are over‑ or under‑represented, leading to misleading conclusions.

Non-response bias

Occurs when the opinions of respondents differ meaningfully from those who did not respond. Low response rates often mean only people with strong opinions participate, skewing results.

Measurement Bias

Arises when the data‑collection method consistently over‑ or under‑represents certain characteristics. This can happen through poor question wording or limited answer choices.

Leading questions

A question that guides or prompts respondents toward a particular answer, often confirming a built‑in assumption.

Loaded questions

A question that contains an assumption, forcing agreement regardless of the answer.

Response bias

Occurs when respondents alter their answers to influence results, avoid embarrassment, or please the questioner.