The BASK expects HR professionals to read charts and to know what a chart can and cannot prove.
The named chart types
The Analytical Aptitude key concept interpretation of graphs and charts names bar charts, line graphs, scatterplots and histograms.
That heading is distinct from:
- Data analysis techniques: data cleansing, data mining, visualization, big data analysis, statistical analysis, predictive analysis, data and algorithm quality.
- Sources of data: a separate key concept on where data comes from.
- Business intelligence techniques and tools: a Business Acumen heading, not an Analytical Aptitude one.
Which chart for which job
- Scatterplot: plots paired values of two variables to show how they relate. Exit survey scores against commute distance across 400 employees is a scatterplot.
- Histogram: the distribution of one variable.
- Bar chart: compares values across categories.
- Line graph: normally shows change over time in one series.
The trap is choosing the line graph or bar chart because they are more familiar. If the stem says "relationship between two variables" or gives paired values, it is the scatterplot.
What a chart does not prove
Correlation is listed under basic statistics and measurement, and it is the idea that cautions against causal conclusions. Engagement scores and sales per employee rising together last quarter is a correlation; it does not by itself establish that engagement produced the sales increase.
- Reliability concerns consistency of measurement, not the inference.
- Validity concerns whether an instrument measures what it is meant to measure.
- Data cleansing is data preparation and has nothing to do with the inference being drawn.
Carry this in: when a manager reads causation off two lines that moved together, the concept you name is correlation.