8 Data Analysis

Question 1

Key weaknesses in data analysis answers

Answer 1

• restating facts of numbers without applying them to the context of the question.
• failing to use additional information given (results in generic answers)
• interpreting figures / results in isolation
• focusing on a narrow range of measures
• failing to use numerical analysis to support the rest of the answer
• failing to explain trends in the data by identifying cause and effect relationships
• failing to achieve a reasonable balance between numerical and descriptive analysis
• failing to understand the impact of accounting standards

Question 2

Recommended approach

Answer 2

1. Review scenario and requirements
2. Decide what analysis is appropriate
3. Produce the calculations
4. Interpret the analysis
5. State the additional information required

Question 3

What How Why When So What analysis

Answer 3

What - look at what has happened eg revenue has increased by 21%

How? - seeks to identify the reasons for the what - price rises and / or volume of sales

Why - look for the underlying causes of the how element - maybe additional features built into product. Why is more important than What. Look for the links between cause and effect.

When - if you’re assessing the impact of changes in strategy over time important to know when changes occurred. Eg if price and volume changes occurred half way through the year the increase in revenues may be limited to say 10.5% but changes more significant in next year’s figures.

So What? The next stage is to ask ‘ so what are the consequences of our analysis for deciding on the future business strategy. Eg what are the consequences for profits, have competitors responded? What might make next year different?

Question 4

Reliability of data analysis

Answer 4

Professional scepticism!

• the data may have been intentionally manipulated - trying to get new finance? trying to sell the co?
• the data may have been analysed accurately but the presentation may have been designed to mislead
• there may be intentional or unintentional bias inherent int he data meaning it is not representative of the population being analysed

Question 5

Selection bias - 5 types

Answer 5

Self selection bias - individuals selecting themselves (eg only those with extremely positive or negative views)

Survivorship bias - only items that survived some previous event are selected (eg staff satisfaction surveys of staff still employed after one year)

Observer bias - researchers allowing their assumptions to influence observations (placebo effect)

Omitted variable bias - a variable being omitted from a data model may lead to the cause of change in one variable being incorrectly attributed to another variable eg increase in sales being attributed to advertising spend when it is in fact caused by increase in competitors price

Cognitive bias - different people may interpret the data in different ways as a result of their own background, experiences and beliefs. Includes confirmation bias eg Brexit debate - finding arguments to support your existing opinions.