Data integrity and issues in transaction processing systems for HSC Information Processes and Technology
“Describe data integrity and concurrency control in transaction processing systems, and the social and ethical issues they raise, including bias in data, the changing nature of work and the move to online and data warehousing”
Data integrity means the stored data is accurate, complete and consistent. A TPS protects it with validation on entry, and with concurrency control so simultaneous transactions do not corrupt shared data. Locking stops two transactions changing the same record at once, preventing problems such as one update overwriting another. The issues a TPS raises include bias built into the data and rules, the changing nature of work as automation replaces manual processing, and the trend toward online and real time processing feeding data warehouses that are mined for patterns, which intensifies privacy and accuracy concerns.
Jump to a section
Legacy course. This page covers HSC Information Processes and Technology (IPT), which NESA has replaced. The last IPT HSC exam was held in 2024; current students study Enterprise Computing 11-12 or Software Engineering 11-12 (first HSC exams in 2025). See the IPT hub for what carries over, or our HSC Software Engineering hub. Use this page as background only.
What this dot point is asking
NESA wants you to describe how a transaction processing system (TPS) keeps its data correct when many transactions happen at once, and the social and ethical issues a high volume transaction system raises. Data integrity and concurrency control are the technical core; bias, the changing nature of work and emerging trends such as online processing and data warehousing are the issues.
The answer
Data integrity
Data integrity is the property that stored data is accurate, complete, consistent and current. A TPS protects integrity at entry through validation (range, type, presence and check digit checks) and verification, and during processing by ensuring every transaction is applied fully or not at all. If a transaction would leave the data in a half-updated, inconsistent state, the system rolls it back so the master file stays correct.
Concurrency control
A busy TPS processes many transactions at the same moment, and several may touch the same record, for example two clerks selling the last item in stock. Without control, simultaneous updates can interfere: one transaction can overwrite another's change, so a sale is lost or stock goes negative. Concurrency control prevents this, most commonly through locking, which gives one transaction exclusive use of a record until it finishes, making others wait. This keeps shared data consistent at the cost of some delay.
The classic concurrency problem is the lost update: two transactions read the same balance, both add to it, and the second write overwrites the first, so one deposit vanishes. Locking the record while it is being updated is what prevents this.
Bias in data and rules
A TPS encodes rules about who gets approved, charged or flagged, and it accumulates data that reflects past behaviour. If the rules or historical data carry bias, the system applies it automatically and at scale, for example consistently disadvantaging a group through a credit rule. Because the processing is automated and trusted, the bias is hard to see and easy to perpetuate, which makes fairness an ethical responsibility of the system's designers.
The changing nature of work
Transaction processing has been heavily automated: self-service checkouts, online banking and automated ordering replace tasks once done by clerks and tellers. This changes work by removing routine processing jobs, shifting remaining staff toward exception handling and customer service, and demanding new skills. The benefits of speed and lower cost are real, but so is the displacement of workers, and a balanced answer holds both.
Online processing and data warehousing trends
The trend is away from overnight batch toward online, real time processing, so data is current the instant a transaction occurs. The transactions captured are increasingly fed into data warehouses, large stores combining historical transaction data, which are mined for patterns to support decisions. This makes the TPS not just an operational system but a source of strategic information, and it intensifies privacy and accuracy concerns because vast, detailed records of behaviour are now retained and analysed.
Tying integrity to the issues
The technical and ethical sides connect: poor integrity produces wrong data, and when that data is mined or used to decide about people, the errors and biases cause real harm. Strong validation, concurrency control and auditing are therefore both technical necessities and ethical safeguards.
Confusing data integrity with security. Integrity is about data being accurate and consistent; security is about preventing unauthorised access. A TPS needs both.
Ignoring concurrency. In a high volume TPS, simultaneous transactions are the norm; without locking or similar control, shared data corrupts.
Describing the lost update wrongly. It happens when two transactions read then write the same value and one overwrites the other; locking during update prevents it.
Giving only the upside of automation. The changing nature of work includes job displacement and new skill demands, not just efficiency.
Treating data warehousing as out of scope. The move to online processing feeding warehouses for mining is a named current trend and a source of privacy issues.
Exam-style questions
Questions in the style of NESA exam questions on this dot point, each with a worked answer. They are written by ExamExplained unless tagged "Past paper"; the year shows the paper a question is modelled on.
2019 HSC-style3 marksDefine bias in data collection. Support your answer with an example.Show worked answer →
For 3 marks give a clear definition and an example that demonstrates it.
Definition: bias in data collection is a systematic distortion in how data is gathered that causes the collected data to unfairly favour, over-represent or under-represent particular outcomes or groups, so it does not accurately reflect reality.
Example: a survey about a shopping centre conducted only on weekday mornings collects responses mainly from retirees and parents not at work, under-representing full-time workers. The results are biased because the sample is not representative of all shoppers.
Another acceptable example is a leading question (for example "Don't you agree our service is excellent?") that pushes respondents toward a particular answer. Markers reward a definition that mentions systematic distortion or unrepresentative data, plus a relevant example.
2020 HSC-style4 marksCustomers use an online concert ticket booking system to choose seats, enter their details and pay. Explain the importance of data validation in this booking system. Include examples of data validation methods that this system would use.Show worked answer →
For 4 marks explain why validation matters here and give named validation methods.
Importance: data validation checks that entered data is reasonable and complete before it is processed and stored. In a ticket booking system it prevents errors that would cause failed payments, wrong bookings or corrupted records, protecting data integrity and giving customers a reliable transaction.
Validation methods this system would use:
Presence (required field) check - ensures essential fields such as name, email and card number are not left blank.
Range check - ensures the number of tickets is within an allowed range (for example 1 to 8).
Type/data type check - ensures numeric fields (card number, quantity) contain only digits.
Format check - ensures the email address and card number match the expected pattern.
Markers reward explaining the purpose (catch errors, protect integrity) plus at least two correctly named validation methods applied to the booking fields.
2021 HSC-style6 marksA department store uses a transaction processing system to keep track of its sales, and these transactions are analysed monthly. Explain how the use of data warehousing would benefit this department store.Show worked answer →
For 6 marks explain what a data warehouse is and develop several benefits for the store.
What it is: a data warehouse is a large central store that collects and integrates historical transaction data from the TPS (and other sources) specifically for analysis, separate from the operational system.
Benefits to the store:
Better decision making. Months and years of sales data in one place let managers spot trends (seasonal demand, best sellers) to guide stock and marketing decisions.
Faster analysis without slowing operations. Because analysis runs on the warehouse rather than the live TPS, monthly reporting does not slow down day-to-day sales processing.
Integrated, consistent data. Data from different stores or systems is cleaned and combined, so reports are based on consistent, reliable information.
Supports data mining and OLAP. The warehouse is the foundation for techniques that find patterns (for example which products sell together), adding further value.
Markers reward a correct description of data warehousing plus several developed benefits tied to the store's monthly analysis.