Study guide: Interpret Data Summaries and Privacy Tradeoffs
For values[1,2,9], calculate the mean and compare it with the median. Then assess a claim that deleting names from detailed unique rows guarantees anonymity.
Ideas to explain in your own words
- A mean is the total divided by the number of values. It summarizes a dataset without preserving every observation. In a small dataset an unusually large value can pull the mean upward. Use the original list as well as the summary before making a statement about typical observations.
- The median is the middle value of an ordered list with an odd number of entries. Ordering is necessary, but it does not mean the original temporal order was meaningless. Here the question concerns only the median; every list contains five recorded numeric observations.
- Aggregation combines records into summaries. Removing names can reduce exposure but does not guarantee anonymity. A very small group or a unique combination of attributes can still reveal information. The task uses fictional records and evaluates claims, without processing anyone’s real personal data.
- A dataset reflects how and when information was collected. A convenient sample may omit people or conditions, and a change in definitions can invalidate comparison. A data claim should preserve population, date and measurement meaning rather than treating a summary as universally representative.
Useful vocabulary
- mean
- Total divided by the number of recorded values.
- median
- A middle value after ordering.
- aggregation
- Combining records into a summary.
Watch your reasoning
- The mean need not equal a value that any item actually has.
- Do not select the third item from an unsorted original list.
- Name removal alone does not establish that reidentification is impossible.
- More decimal places do not repair biased sampling or incompatible definitions.
A worked review example
Paper Python 3 expression: sum([2, 4, 6])/len([2, 4, 6]). What value results? The supported answer is 4.0. Tracing the expression under Python 3 rules gives 4.0. Add the values and divide by the number of supplied values. The mean need not equal a value that any item actually has. A mean is the total divided by the number of values. Only one club supplies responses. Which limit is present in the record? The supported answer is Limited sampled group. Check the source group, time and measure before generalizing or comparing. More decimal places do not repair biased sampling or incompatible definitions. A dataset reflects how and when information was collected.
Make a practice plan
- Make a four-row summary, one row for each lesson, with its computing idea and one worked decision.
- Explain one tempting mistake using the actual supplied example.
- Complete the new unit review before consulting its explanations.
- Try the transfer task on paper and compare your evidence against the model response.