Data Collection Plan Homework
Eva Chan 
Chapter13_ASQ.pptDS-624 Quality Management
Chapter 13
D. Collecting and Summarizing Data
DS-624 Quality Management
Agenda
What are the 2 types of data?
What is continuous data?
What is discrete data?
What are the advantages of continuous data?
Why is a data collection plan important?
What are the methods to collect data?
What can we do to minimize data collection errors?
Describe the types of Sampling
What is the purpose of descriptive statistics?
Describe the different graphical methods
Types of Data and Measurement Scales
What are the 2 types of data?
- Quantitative data is grouped into 2 types: continuous (also called variable) and discrete (also called attributes).
What is continuous data?
- Continuous data result from the measurement on some continuous scale such as length, weight, temperature, and so on. These scales are called continuous because between any two values there are an infinite number of other values.
Types of Data and Measurement Scales
What is discrete data?
- Discrete data result from counting the occurrence of events. Example might include the number of valve that leaks
What are the advantages of continuous data?
- Control charts based on continuous data a more sensitive to process changes than those based on discrete data.
Data Collection
Why is a data collection plan important?
- Data collection can be very expensive if the data collection is not planned and effectively implemented.
- Answering some basic questions before actually starting to collect the data such as what, why, where, when, who, and how, can help make planning the data collection more effective.
Data Collection
What are the methods to collect data?
- Surveys
- Face-to-face interviews
- Focus groups
- Mystery shopping
- Customer feedback
- Automatic data capture
- Manual data capture
Techniques for assuring data accuracy and integrity
What can we do to minimize data collection errors?
- Have a carefully constructed data collection plan following 5W1H
- Maintain a calibration schedule for data collection equipment
- Conduct repeatability and reproducibility (R&R) studies on the measurement system
- Use appropriate statistical tests to remove outliers
- If data are transmitted or stored digitally, use an appropriate redundant error correction system.
- Provide clear and complete instruction and training for collection, transformation, analysis, and interpretation.
Types of Sampling
Describe the types of Sampling
- Random Sampling: Every sample randomly picked from the lot has equal probability of getting picked. If effectively administered, Sampling can save money for the organization.
- Sequential Sampling: It is used in destructive testing and reliability testing applications where higher cost is involved in testing the unit.
- Stratified Sampling: Where there is a mixture of parts from different machines, different streams, different raw material lots or different process settings., there is no homogeneity of the lot. Hence, random sampling, will not yield the right results. It will be more effective to stratify the lot based on the criteria.
Descriptive Statistics
What is the purpose of descriptive statistics?
- The purpose of descriptive statistics is to present data in a way that will facilitate understanding
Graphical Methods
Describe the different graphical methods
| Name | Purpose | Application | Interpretation | Ease of Use |
| Tally | Provides a quick tally of total quantity and by class interval. Provides visual ideas of the distribution shape | Used to count defect quantity by type, class and/or category | Tally mark concentration and spread roughly indicate distribution shape. | Very easy to create and interpret |
| Frequency Distribution | Provides a pictorial view of numerical data about location and spread. | Especially useful if tally column cells have a large number of marks. | Concentration of data is seen as a pic, and spread of data is demonstrated by the width of the curve. Thinner distribution indicates lesser variation. Distribution can be unimodal (with one peak), bimodal (with 2 peaks) or multimodal (multiple peaks) indicating a mixture of populations. | Not so easy to create but easy to interpret |
| Stem-and-leaf plot | Provides a pictorial view of minimum, maximum, median, and interquartile range in one graph. | Useful to quickly identify any repetitive data within the class interval | If data values within cells are not fairly evenly distributed, measurements errors or other anomalous conditions may be present. | Easy to create but difficult to interpret |
Graphical Methods
| Name | Purpose | Application | Interpretation | Ease of Use |
| Box-and-whisker plot | Provides a pictorial view of minimum, maximum, median, and interquartile range in one graph. | Provides more information than distribution plot but easier to interpret. Outliers are easily identified on the graph. | If the location of the center of the box is right in the middle, the data may be normally distributed. The data points outside the whiskers indicate outliers. Unequal whiskers indicate skewness of the distribution. | Easy to create and interpret |
| Scatter Diagram | Detects possible correlation or association between two variables, or cause and effect. | Used for root cause analysis, estimation of correlation coefficient,. | To estimate correlation the relationship has to be linear. If the data flows upward from left to right then it is positive correlated. If the date flows downward from left to right the relationship is negatively correlated. If the data are spread about the center whit no inclination, then there may not be any correlation. | Easy to create and interpret |
| Run chart | Provide a visual indicator of any nonrandom patterns | Used when real-time feedback is required for variable data | Patterns like cluster, mixture, trend, and oscillation are spotted based on the number of runs above or below the mean or median. | Easy to create and interpret |
