De-identifying survey and interview data with participant codes

Name:

Date:

Class:

Quick Questions

Shade in the A or B circle for each question.

1. A student replaces each survey respondent's name with a code such as P01 before analysing the data. What is this called?

A De-identifying the data using participant codes

B Encrypting the data so nobody can read it

2. Which interview transcript line is better de-identified?

A Mia Chen: I think the app is too slow.

B P03: I think the app is too slow.

3. After replacing names with codes, a dataset still lists each person's school and suburb. Is it fully de-identified?

A No, school and suburb are indirect identifiers and should be removed or generalised

B Yes, once names are gone nobody can be identified

4. Which is an indirect identifier?

A The suburb and school of a respondent

B The code P07 assigned to a respondent

5. Where should the key list linking codes (P01, P02...) to real names be stored?

A In an extra column in the same dataset

B Separately from the de-identified dataset, with restricted access

Match the Pairs

Draw a line connecting each item on the left to its match on the right.

Left column		Right column
1P01		AThe name, school and suburb are still in the dataset, so the participant can still be identified.
2Key list: P01 = Sam Tran		BThe key list is stored in the same file as the data, so anyone with the file can re-identify participants.
3Indirect identifiers such as school and suburb are removed so the person cannot be worked out		CP03: I mostly use my phone to study.
4Each participant gets a code in place of their name in the transcript		DA separate record linking each code to a real person, stored apart from the dataset
5P01, Sam Tran, Year 11, Glen Waverley SC		EDelete columns: School, Suburb
6Survey_results.xlsx (sheet 2: P01 = Sam Tran, P02 = Mia Lee)		FA participant code used in the dataset and transcripts in place of the person's name

Definitions

participant code

Fill in the Blanks

Complete each line of code by filling in the missing parts.

SELECT _______________, answer FROM survey;

ALTER TABLE survey ____ COLUMN suburb;

CREATE TABLE _______ (code, name);

participantCode ← "P0" _ id

UPDATE survey SET school = ____;

Spot the Error

Shade in the letter of the line that contains an error.

Question 1

A FOR each survey response

B code = next participant code

C DELETE name but keep suburb

D SAVE code with the answers

Question 2

A Header: code,year,answer

B Row 1: P01,Year 10,Agree

C Row 2: Mia,Year 11,Disagree

Question 3

A Assign codes P01, P02, P03

B Make a key list: code to name

C Swap names for codes in data

D Keep key list in the data file

Question 4

A code = name[:3] + str(n)

B record['code'] = code

C del record['name']

D del record['school']

Question 5

A Raw: Ava, Lowood, Year 10

B De-id: P01, Lowood, Year 10

C Key: P01 = Ava (stored apart)

Answer Key

De-identifying survey and interview data with participant codes

Section A: Quick Questions

A — De-identifying the data using participant codes

B — P03: I think the app is too slow.

A — No, school and suburb are indirect identifiers and should be removed or generalised

A — The suburb and school of a respondent

B — Separately from the de-identified dataset, with restricted access

Section B: Match the Pairs

P01 → A participant code used in the dataset and transcripts in place of the person's name

Key list: P01 = Sam Tran → A separate record linking each code to a real person, stored apart from the dataset

Indirect identifiers such as school and suburb are removed so the person cannot be worked out → Delete columns: School, Suburb

Each participant gets a code in place of their name in the transcript → P03: I mostly use my phone to study.

P01, Sam Tran, Year 11, Glen Waverley SC → The name, school and suburb are still in the dataset, so the participant can still be identified.

Survey_results.xlsx (sheet 2: P01 = Sam Tran, P02 = Mia Lee) → The key list is stored in the same file as the data, so anyone with the file can re-identify participants.

Section C: Definitions

participant code — A code such as P01 used instead of a name. Names and indirect identifiers like school or suburb are removed, and the key list linking codes to people is stored separately.

Section D: Fill in the Blanks

SELECT participantCode, answer FROM survey;

ALTER TABLE survey DROP COLUMN suburb;

CREATE TABLE keyList (code, name);

participantCode ← "P0" + id

UPDATE survey
SET school = NULL;

Section E: Spot the Error

C — A suburb is an indirect identifier. Combined with other details it can point to a person. Both the name and the suburb must be deleted: DELETE name and suburb.

C — Every row must use a participant code, not a name. Row 2 should start with P02 so the person cannot be identified.

D — The key list links codes to real people, so it must be stored separately from the dataset. Keeping it in the data file defeats the de-identification.

A — A participant code must not be built from the person's name, because it still reveals who they are. It should be a neutral code like 'P' + a number, e.g. P01.

B — The suburb Lowood is an indirect identifier and must be removed from the de-identified data. The line should be De-id: P01, Year 10.