Quantitative vs Qualitative
Quantitative data is measurable and specific, and therefore it’s easy to chart and graph. Quantitative data is gathering data based on verifying a research question through the use of statistics and data that is largely numerical.
qualitative data provides a more in-depth understanding. It’s focused on more open and more descriptive answers, often asking for open-ended questions rather than a measurable data like tallying a yes or no.
Surveys, focus groups, observations are quantitative (numerical).
Qualitative data is about qualities or attributes, and it is much harder to measure than quantitative data. Usually these data are open-ended like why do you like red?
Focus groups (text), observations are qualitative when looking at the descriptions and themes, linking them together.
Quality of data
One aspect of identifying relevant data from a given data set is usable. To be usable, data must be accurate, relevant, free from bias, and reliable.
Relevance
To produce usable information, data must be relevant. For example, if a computing department is evaluating PC-only software, then surveying people who only use an Apple device is an example of not relevant data.
Accuracy
Data that is collected must be accurate, otherwise the insights will be poor and decisions based on this data could cause integrity issues.
Ensure data is accurate otherwise the insights will be poor and decisions based on this data could cause integrity issues.
Free from bias
Bias can easily creep into data and make the information processed from it unreliable. Several influences can lead to the introduction of bias into data; these include vested interests, timing, and small sample sizes.
Consider any social factors that might cause any bias in any data collection methods. Bias can make data unreliable and unfair. Bias can occur from vested interests, timing, smaller sample sizes, biased sorting, bias in graphic representations. For example, you cannot get opinions on which phone is better if you just ask Apple users.
| Vested interest | If the participant or the interviewer have a prior interest in a certain outcome. |
| Timing | The timing of data may introduce biases too. For example, opinions on Russia can be impacted by their war on Ukraine compared to opinions on Russia before their war. |
| Small sample sizes | A small sample size will result in less representation from certain groups. For example, asking for opinions on Chinese cuisine will limit the amount of representation from other types of people. |
| Bias through sorting | Although it is nearly unavoidable, sorting can create biases. For example, alphabetically sorting a list of names will place names at the top - making their names the first thing people will see. |
| Bias in graphic representations | Typefaces, colour choices, sizes and different types of graphic representations can make certain data appeal more to others - again, creating biases. |
Reliability
Ensure the data is reliable.
Timing
The timing of the data collection may also introduce bias. If the data is recorded decades ago, the data is no longer useful and relevant. Ensure the data is relevant at the current time.
Usually human error is the biggest source of data inaccuracy.