Data Collection and Analysis
Gathering and Making Sense of Data
Good analysis starts with disciplined collection. Decide in advance whether data are quantitative (counts, measurements, scores) or qualitative (interviews, observations, open responses), and design instruments - checklists, rubrics, surveys, logs - that produce consistent records. Use calibration, repeated trials, and clear operational definitions to improve reliability. Record raw data faithfully; do not "fix" outliers without justification.
Concepts to Know
- Descriptive statistics: mean, median, mode, range, and standard deviation as measures of center and spread.
- Data displays: tables, bar and line graphs, scatterplots; choose the display that fits the data type.
- Outliers and their effect on the mean versus the median.
- Qualitative coding: grouping responses into themes with defined criteria.
- Measurement error and bias in instruments and observers.
Traps in Questions
- Reporting a mean when data are skewed and the median is more informative.
- Misreading a graph with a truncated axis or inconsistent scale.
- Treating a small difference as meaningful without considering spread.
- Cherry-picking favorable data points and discarding the rest.
- Confusing percentages of different bases when comparing groups.
Worked Example / Rule of Thumb
Rule of thumb: always report center with spread and check the shape first. Example: reaction times of 12 participants are mostly 300–340 ms, but one is 900 ms. The mean (about 375 ms) is pulled upward by the outlier; the median (about 320 ms) better represents typical performance, and the large range signals variability worth discussing rather than hiding.