Research Methods Chapter 14: Data Collection

Study Notes

Data collection is a precise, systematic procedure for gathering relevant information for research.
A researcher, after the research proposal is approved, collects data according to the research plan.
The success of any research depends on the quality of collected data, its collation, testing, analysis, and interpretation.
Data refers to the kinds of information obtained from research participants.
Different types of data collection methods exist including cross-sectional, retrospective, and prospective.
Primary data collection involves personally gathering data from respondents (e.g., interviews, questionnaires).
Secondary data collection uses data previously collected for other purposes (e.g., medical records, surveys).
Steps in data collection include explaining and clarifying the measures and tasks associated, and administering and describing the methods used.

Data collection methods depend on the research purpose, method, and tools used.
Cross-sectional data is collected at a single point in time.
Retrospective data (ex post facto studies) involves collecting data on past events.
Prospective data involves collecting data on future events or conditions

Primary data involves personally collecting data (e.g., interviews, questionnaires).
Secondary data uses pre-existing data (e.g., medical records, surveys).

Using existing data (records, documents).
Using observers' data (actual observation, recording of events).
Non-participating observer: Observer doesn't partake in the subjects' environment.
Participating observer: Observer works within the subjects' environment.

Structured observation: Researcher already has knowledge of the phenomenon of interest; structured observation employs checklists to record observed behaviors.
Unstructured observation: Researcher uses no preconceived ideas and aims to describe all behaviors observed with careful attention.

Overt, non-participant observer: Observer identifies herself and her purpose to participants.
Covert, non-participant observer: Observer does not disclose identity or purpose to participants.
Overt, participant observer: Observer openly participates with the subjects while observing their behaviors.
Covert, participant observer: Observer participates but does not disclose their observation efforts.

Research variables are defined before they are measured.
Measurement (in research) is the process of assigning numerical value to variables according to rules.
Measurement levels include nominal, ordinal, interval, and ratio.
Nominal scale classifies variables without ordering.
Ordinal scale ranks variables in order.
Interval scale ranks variables with equal intervals, but doesn't have a true zero.
Ratio scales have a true zero point, and can be meaningfully added or subtracted.
Measurement errors may occur, and researchers use strategies to minimize errors. These include:
Reliability (consistency of a measure)
Accuracy (truthfulness of a measurement)
Content validity (adequacy of the content of a measure).
Face validity (measure looks as if it measures the intended construct).
Criterion-related validity (correlation with an external criterion).
Construct validity (testing theoretical constructs).
Validity & Reliability in qualitative research.
Reliability in quantitative research.
Strategies to minimize measurement errors.

Statistics is used in nursing research to summarize and interpret numeric data.
Descriptive statistics describes the sample studied (e.g., mean, median, mode).
Inferential statistics uses sample data to draw conclusions about the larger population.
Statistical tools for treatment of data (e.g., frequency distribution, percentage distribution, means, standard deviation, T-test, ANOVA, regression) are used in quantitative research
Common problems in measurement includes inaccurate measurement tools, sampling problems, and issues related to variability in data collection.