Work Sampling: A Statistical Tool for Efficient Management

Introduction

In today’s business world, efficiency and productivity are essential for success. To achieve these goals, companies need to understand how time is used in their operations. Work measurement provides the necessary tools to analyze and optimize work processes, and among the most versatile and accessible techniques is work sampling.

Work sampling, also known as activity sampling, is a statistical technique that allows you to determine the percentage of time dedicated to a specific activity or set of activities within a given period.

Unlike continuous observation, which requires exhaustive dedication and can be costly, work sampling is based on observing activities at random moments, making it an efficient and low-cost tool.

Why Use Work Sampling

In complex work environments, where multiple activities are carried out simultaneously, it is practically impossible to continuously observe and record everything that happens. Continuous observation is not only costly in terms of time and resources but can also interfere with the normal work rhythm of employees, generating biased data.

Work sampling emerges as a solution to this challenge. By observing activities at random moments, a representative sample of the work being performed is obtained, providing an accurate view of time distribution without the need for continuous observation.

Example: An appliance factory has an assembly line where various operations are performed, such as component placement, wiring, functional testing, and packaging.

Continuously observing all these activities would be an impossible task. However, through work sampling, random observations can be made throughout the day to determine the percentage of time the line is operating, the percentage of time it is stopped, and the causes of the stoppages.

This information allows for identifying bottlenecks in the process, optimizing production planning, and improving the overall efficiency of the assembly line.

Work sampling is a valuable tool for companies seeking to improve efficiency and productivity. By providing an accurate view of time distribution, it allows for identifying areas for improvement, optimizing planning and production control, and making informed decisions to achieve business objectives.

Work Sampling: Principles and Representativeness

Work sampling is based on statistical principles, specifically the law of probabilities. This law establishes that the greater the number of observations, the greater the precision of the results obtained. In other words, the larger the sample, the more faithfully it will represent the “population” or “universe” being studied.

To illustrate this concept, imagine a bag full of marbles of different colors. If we randomly extract a handful of marbles and count how many are red, we will get an approximate idea of the proportion of red marbles in the bag. However, if we repeat this process several times and with a larger number of marbles, our estimate of the proportion of red marbles will be more precise.

In work sampling, the “population” refers to the set of activities that take place in a given period of time. By making random observations, we are extracting a sample from this population. The key to obtaining reliable results is to ensure that the sample is representative, meaning it faithfully reflects the actual distribution of time among different activities.

Example: In a study of office work, it is observed that on a given day, 30% of observations correspond to employees attending to customers, 50% to working on computers, and 20% to performing other activities. If the sample size is large enough and the observations have been made randomly, we can trust that these percentages are representative of the actual distribution of time in the office.

Randomness in the selection of observations is crucial to avoid biases in the results. If observations are made at specific times or with a predetermined frequency, they may not reflect the true distribution of time. For example, if we always observe workers early in the morning, we may overestimate the time dedicated to work preparation and underestimate the time dedicated to other activities.

Work sampling is a powerful tool for obtaining accurate information about time distribution in organizations. By understanding the statistical principles that support it and ensuring the representativeness of the sample, companies can make informed decisions to improve efficiency and productivity.

Complete Guide to Conducting a Work Sampling Study

1. Define the Objective of the Study

• What do you want to measure? Precisely determine which aspect of the work will be analyzed.

Examples of objectives:

• Identify and quantify unproductive time in a production line.
• Establish the proportion of time dedicated to different office activities.
• Evaluate the utilization of a specific machine.
• Determine the frequency of errors in an assembly process.

Example: The goal is to determine the percentage of time that an injection molding machine is stopped due to lack of material.

2. Select the Work to be Studied

• Choose your focus: Decide whether the study will focus on a machine, a group of workers, a process, or an entire department.

• Consider the scope: Define the time period that the study will cover (days, weeks, months).

Example: The injection molding machine is chosen as the object of study, and a two-week observation period is defined.

3. Break Down the Work into Elements

• Divide and conquer: Break down the work into specific activities that are easy to identify and observe.

Examples of elements:

• Machine running/stopped.
• Worker working/waiting.
• Type of activity performed by the worker (cutting, drilling, filing, etc.).
• Machine status (operational, under repair, without material, etc.).
• Type of error in the process (missing, defective, etc.).

Example: The work elements are defined as: machine running, machine stopped due to lack of material, machine stopped due to other causes.

4. Determine the Sample Size

The sample size in work sampling is a critical factor that affects the precision and reliability of the results. A sample that is too small can lead to erroneous conclusions, while a sample that is too large can result in a waste of time and resources.

The sample size needed for a work sampling study depends on two main factors:

• Confidence level: As explained earlier, the confidence level represents the certainty you want to have in the results. A higher confidence level requires a larger sample size.

• Margin of error: The margin of error is the amount of variation you are willing to accept in the results. A smaller margin of error requires a larger sample size.

Confidence and precision: Choose the confidence level (e.g., 95%) and the margin of error (e.g., 5%) you want for your results.

Calculation methods:

Statistical method:

This method uses a mathematical formula that takes into account the confidence level, the margin of error, and an initial estimate of the proportion of time dedicated to the activity being studied.

Use the following formula to calculate the sample size (n):

n = p * q / (σp)^2

Where:

• n: sample size
• p: percentage of inactive time (from preliminary observation)
• q: percentage of running time (from preliminary observation)
• σp: standard error of the proportion (calculated considering the confidence level and margin of error).

Nomographic method:

This method uses a special chart called a nomogram that allows obtaining the sample size quickly and easily. The nomogram presents scales for the confidence level, margin of error, and sample size, and is used by drawing a straight line between the scales to obtain the result.

Data analysis software:

Digital tools can be used to facilitate the calculation of the sample size.

Example: A company wants to conduct a study on efficiency in a call center.

They want to determine the percentage of time that operators dedicate to answering calls, with a 95% confidence level and a 5% margin of error.

Using the statistical or nomographic method, it is calculated that 385 observations are needed. This means that the activities of the operators will need to be observed at 385 random moments during the study period.

The choice of the sample size calculation method will depend on the availability of tools and the analyst’s preference. In any case, it is important to understand that the sample size has a direct impact on the precision and reliability of the study results.

By balancing precision with efficiency, companies can obtain valuable information about their work processes without incurring excessive costs.

5. Conduct Random Observations

• Randomness as a tool: Use a table of random numbers to select the moments when observations will be made during the study period.

• Divide the time: Divide the observation period into equal time intervals (e.g., 10 minutes) and assign a number to each interval.

• Select randomly: Use the table of random numbers to choose the intervals in which observations will be made.

Example: The 8-hour workday is divided into 48 intervals of 10 minutes. Using the table of random numbers, 385 random moments are selected within these intervals to make the observations.

6. Record the Observations

• Record sheet: Design a record sheet with the identified work elements.

• Simple recording: In each observation, note the element that is being performed at that moment.

• Electronic devices: Electronic data collection devices can be used to automatically record observations.

Example: On the record sheet, it is noted whether the injection molding machine is running, stopped due to lack of material, or stopped due to other causes in each of the 385 observations.

7. Analyze the Results

• Quantify the observations: Sum the number of times each element was observed.

• Calculate percentages: Divide the number of observations for each element by the total number of observations and multiply by 100 to obtain the percentage of time dedicated to each activity.

Example: Of the 385 observations, the machine was running 250 times, stopped due to lack of material 85 times, and stopped due to other causes 50 times. This means that productive time is 65%, unproductive time due to lack of material is 22%, and unproductive time due to other causes is 13%.

8. Take Corrective Measures

• Interpret the results: Analyze the information obtained in relation to the objective of the study.

• Identify areas for improvement: Based on the results, determine the causes of unproductive time or inefficiencies in the process.

• Implement solutions: Design and implement corrective measures to improve efficiency and productivity.

Example: It is confirmed that lack of material is the main cause of unproductive time. Work is done with the purchasing department to improve the inventory control system and ensure the availability of materials for the injection molding machine.

Additional Recommendations:

• Team training: It is important to train employees who will participate in data collection on how to make observations accurately and consistently.

• Clear communication: It is essential to communicate to employees the objective of the study and its importance for improving efficiency.

• Work pace analysis: If the objective is to evaluate the work pace, a pace rating scale can be used to adjust the results of the study.

• Use of digital tools: Data analysis tools can facilitate the collection, analysis, and visualization of the study results.

Establishing Confidence Levels

Work sampling, being based on the observation of a representative sample, generates probabilistic results. This means there is a margin of error associated with the study’s conclusions. To quantify this margin of error and express certainty in the results, confidence levels are used.

Confidence levels are based on the normal distribution curve, a graphical representation of the frequency with which observations are distributed around the mean. In a normal curve, most observations are found near the mean, while observations further away are less frequent.

A 95% confidence level, for example, indicates that if we repeat the study many times, 95% of the time the results will be within a determined margin of error. In other words, there is a 95% probability that the true proportion of time dedicated to an activity is within the interval defined by the margin of error.

Example: In a study on equipment utilization in a mechanical workshop, a 95% confidence level and a 5% margin of error are obtained. This means we can trust that 95% of the time, the true proportion of time that the equipment is in operation is within a range of ± 5% of the value observed in the study.

Common confidence levels in work sampling are 95% and 99%. The choice of confidence level will depend on the importance of precision in the results and the consequences of a possible error. In general, a higher confidence level requires a larger sample size.

Understanding confidence levels is essential for properly interpreting the results of work sampling. By expressing certainty in the results, confidence levels allow companies to make informed decisions based on statistically sound information.

Conducting Random Observations

The key to obtaining reliable results in work sampling lies in the randomness of the observations. This means that each moment of observation must be selected at random, without any type of bias or predetermined pattern.

Tables of random numbers are an essential tool for ensuring randomness in work sampling. These tables contain a series of digits generated randomly, which means that each digit has the same probability of appearing.

To use a table of random numbers in work sampling, the following procedure is followed:

• Divide the study period into time intervals: The workday, work shift, or observation period is divided into equal time intervals. For example, if studying an 8-hour workday, it can be divided into 48 intervals of 10 minutes.

• Assign numbers to the intervals: A unique number is assigned to each time interval.

• Randomly select the observation intervals: The random numbers from the table are used to choose the intervals in which observations will be made. For example, if the first random number from the table is 23, the first observation will be made in interval 23, which corresponds to a specific moment within the workday.

Example: A supermarket wants to conduct a study on waiting time at the checkout counters. The workday is divided into 15-minute intervals, and a table of random numbers is used to select 200 observation moments. At each selected moment, the number of customers waiting in line is noted.

Tables of random numbers are a simple but effective tool for ensuring randomness in work sampling. By using this technique, companies can obtain a representative sample of activities and avoid biases in the results.

Alternatives to tables of random numbers:

• Random number generation software: There are computer programs that automatically generate random numbers.

• Systematic sampling techniques: In some cases, systematic sampling can be used, selecting observations at fixed intervals, as long as the chosen interval does not coincide with any natural cycle of the work.

The choice of method for selecting observations will depend on the analyst’s preferences and the availability of tools. The important thing is to ensure that the process is random and that the sample is representative of the work being studied.

Implementing the Study: Recording and Observation

With the study objective defined, the sample size calculated, and the observation moments selected, we are ready to implement the work sampling study. This stage focuses on accurate recording and detailed observation of activities.

Preparation of record sheets:

• Clear and concise design: Record sheets should be easy to use and understand.

• Study identification: Include information such as the study name, date, observer’s name, etc.

• Activity categories: List the different activities or work elements that will be observed, leaving enough space to record the observations.

• Flexible format: The format of the record sheet can vary according to the type of study and the analyst’s preferences.

Example: In a study of time dedicated to different tasks in a marketing department, the record sheet could include the following categories:

• Content creation (article writing, graphic design, etc.)
• Social media management
• Data analysis
• Meetings
• Email and internal communication
• Other activities

Conducting the observations:

• Accurate observation: Record the activity being performed at the exact moment of observation.

• Avoid interference: Observations should be discreet and should not interfere with the normal work rhythm of employees.

• Objective recording: Record observations objectively, without value judgments or subjective interpretations.

Example: The observer approaches the workstation of an employee in the marketing department at the randomly selected moment. They observe that the employee is working on creating a graph for a presentation. The observer records this observation on the record sheet under the “Content creation” category.

Calculation of results:

• Sum the observations: Once the study is completed, the observations for each activity category are summed.

• Example: At the end of the marketing department study, 100 observations are counted for “Content creation”, 75 for “Social media management”, 50 for “Data analysis”, etc.

Accurate recording and detailed observation are essential for obtaining reliable results in work sampling. By carefully planning the record sheet and conducting observations objectively, companies can obtain valuable information about the distribution of time in their operations.

Measuring Pace: Normal Pace Work Sampling

So far, we have focused on work sampling as a tool to determine the percentage of time dedicated to different activities. However, this technique can also be used to evaluate the work pace of employees in relation to a “normal” or standard pace.

Normal pace work sampling, also known as normal activity sampling, combines random observation with performance rating. This means that, in addition to recording the activity being performed, the observer also evaluates the speed at which the worker is performing the task compared to a predefined standard pace.

To perform the performance rating, the observer uses a rating scale that assigns numerical values to different levels of work pace. For example, a value of 100 may represent the “normal” pace of a qualified worker, while values above 100 indicate a faster pace and values below 100 indicate a slower pace.

Example: A fruit packaging company wants to evaluate the work pace of employees on a packaging line.

A normal pace work sampling study is conducted, observing the activities of workers at random moments and rating their work pace using a scale from 100 to 133, where 133 represents an optimal pace and 100 represents the normal pace.

At the end of the study, the following results are obtained:

• Observed work time: 400 minutes.
• Number of observations: 200.
• Average work pace rating: 80.

To calculate the normalized work time, the following formula is used:

Normalized work time = Observed work time x (Average pace rating / Standard pace)

Normalized work time = 400 minutes x (80 / 100) = 320 minutes

This means that if the workers had worked at the standard pace throughout the observation period, they would have completed the same amount of work in 320 minutes.

Normal pace work sampling is a useful tool for:

• Setting time standards: Establishing standard times for different tasks as a basis for incentive systems or workforce planning.

• Evaluating efficiency: Identifying workers who are working below the standard pace and providing them with training or support to improve their performance.

• Comparing performance: Comparing the work pace between different workers, teams, or departments.

It is important to note that performance rating is a subjective activity that requires experience and training on the part of the observer. Appropriate rating scales should be used for the type of work being studied, and care should be taken to avoid biases in the evaluation.

Team Efficiency: Group Sampling Techniques

Teamwork is a reality in many work environments. To evaluate the efficiency and productivity of these groups, group sampling techniques, also known as high-frequency sampling, have been developed.

These techniques are characterized by making observations at very brief fixed intervals, which allows obtaining a detailed view of teamwork and analyzing the interactions between group members. They are especially useful for short-cycle work, where individual activities are completed in a few seconds or minutes.

Methodology:

• Definition of work elements: The individual activities that make up the group’s task are identified, as well as the interactions between team members.

• Fixed intervals: A fixed time interval is established for making observations. The interval should be brief enough to capture variations in the work, but long enough for the observer to record the information accurately.

• Recording observations: At each interval, the activity being performed by each group member is observed and recorded on the record sheet.

• Calculation of results: The observations for each activity are summed, and the percentage of time dedicated to each one is calculated.

• Performance rating (optional): A rating scale can be used to evaluate the work pace of the group or each member individually.

Example: A team of operators works on an electronic component assembly machine. The group’s task is divided into the following activities:

• Operator 1: Place the base plate in the machine.
• Operator 2: Insert the electronic components into the base plate.
• Operator 3: Solder the components.
• Operator 4: Perform a functionality test.

A group sampling study is conducted with 30-second intervals. At each interval, the activity being performed by each operator is observed and recorded on the record sheet. At the end of the study, the results are analyzed to identify possible imbalances in the workload, unnecessary waiting times, or areas for improvement in team coordination.

Advantages of group sampling:

• Detailed view of teamwork: Allows analyzing the interactions between group members and identifying areas for improvement in coordination and communication.

• Suitable for short-cycle work: Captures variations in the work accurately.

• Evaluation of team performance: Allows measuring the efficiency of the group as a whole and each member individually.

Limitations of group sampling:

• Requires a greater number of observations: Short observation intervals imply the need for a greater number of observations to obtain reliable results.

• Greater complexity: The analysis of results can be more complex than in individual sampling due to the interactions between group members.

Group sampling is a valuable tool for evaluating the efficiency and productivity of teamwork. By providing a detailed view of group activities, it allows identifying areas for improvement and optimizing work organization to achieve better performance.

Putting Information into Action: How to Use Work Sampling

Work sampling is not limited to being a simple data collection exercise; its true value lies in the application of the information obtained to improve efficiency and productivity in organizations. Let’s look at some key applications:

Identification of unproductive time and areas for improvement

• Analyze the causes of stoppages: Work sampling can reveal the proportion of time that machines or workers are inactive and the reasons for these stoppages. This information is crucial for identifying bottlenecks in processes, optimizing production planning, and reducing unproductive time.

• Example: In a textile factory, work sampling reveals that weaving machines are stopped 20% of the time due to lack of thread. A more efficient inventory control system is implemented to ensure the availability of materials and reduce unproductive time.

Production planning and control

• Balance workloads: Work sampling allows analyzing the distribution of time between different tasks and workers. This information is useful for balancing workloads, avoiding overloading some workers, and ensuring efficient use of labor.

• Example: In a call center, work sampling reveals that some operators are handling a greater number of calls than others. Calls are redistributed more equitably to optimize the time of all operators.

Quality control

• Identify the causes of defects: Work sampling can help identify the stages of the production process where more defects are generated. This information is crucial for implementing corrective measures and improving the quality of the final product.

• Example: In a food factory, work sampling reveals that a high percentage of defective products are generated in the packaging stage. Workers are trained in proper packaging techniques, and stricter quality control measures are implemented.

Setting performance standards

• Establish standard times: Work sampling, combined with pace rating, can be used to establish standard times for different tasks. This information is the basis for workforce planning, setting production goals, and implementing incentive systems.

• Example: In a computer assembly company, work sampling is used to establish the standard time for assembling a computer. This information is used to calculate the number of workers needed to meet production goals and to implement a performance-based payment system.

Applications in different sectors:

Work sampling is a versatile technique that can be applied in a wide range of sectors, such as manufacturing, construction, services, offices, etc. Its flexibility makes it a valuable tool for any organization seeking to improve efficiency and productivity.

Examples of applications:

• Industry: Identification of unproductive time in production lines, evaluation of machinery utilization, quality control.

• Services: Study of efficiency in customer service centers, analysis of waiting time in restaurants, evaluation of productivity in hospitals.

• Offices: Analysis of time dedicated to different administrative tasks, study of efficiency in document processing, evaluation of workload in different departments.

Work sampling is a powerful tool that provides valuable information for decision-making in organizations. By understanding its applications and using the information obtained strategically, companies can improve their processes, increase productivity, and achieve their efficiency and profitability goals.

Structured Estimation: Alternatives for Work Measurement

While work sampling is an effective technique for work measurement, there are situations where it may be less suitable or impractical. In these cases, structured estimation can be used, a technique that combines estimation with data synthesis to obtain approximate work times.

There are two main types of structured estimation:

• Analytical estimation: This method is based on breaking down the work into elements and using standard or synthetic data to calculate the time needed to complete the task. Standard data can come from previous time studies, predetermined time systems, or standard time databases.

• Comparative estimation: This method is based on comparing the work being measured with reference works of known content. Reference works that are similar to the work being studied are selected, and the characteristics of both works are compared to estimate the time needed to complete the task.

Example: An industrial plant needs to estimate the time needed to perform preventive maintenance on a complex machine. Due to the unique nature of the task and the lack of historical data, it is decided to use structured estimation.

- Analytical estimation: The maintenance work is broken down into elements such as component cleaning, lubrication, parts inspection, and filter replacement. Standard time databases are consulted to obtain standard times for each element, and these times are added to obtain an estimate of the total time needed for maintenance.

- Comparative estimation: Similar maintenance work performed on other machines in the plant is identified. The characteristics of these machines are compared with the machine being studied, and the times of the reference works are adjusted to obtain an estimate of the time needed for maintenance of the machine in question.

Advantages of structured estimation:

• Lower cost and time: It is less costly and requires less time than work sampling or time study.

• Useful for non-repetitive work: It is suitable for unique or infrequently performed work, where the investment in a complete time study is not justified.

• Ease of application: It can be applied with minimal training and does not require specialized skills in work study.

Limitations of structured estimation:

• Lower precision: The times obtained are estimates and may have a greater margin of error than times obtained with work sampling or time study.

• Dependence on existing data: The accuracy of the estimate depends on the quality and relevance of the standard data or reference works used.

Structured estimation is a viable alternative to work sampling in certain situations. By combining estimation with data synthesis, companies can obtain approximate work times quickly and efficiently.

Considerations for choosing the work measurement technique:

The choice between work sampling and structured estimation will depend on several factors, such as:

• Nature of the work: For repetitive and short-cycle work, work sampling or time study may be more suitable. For non-repetitive or long-cycle work, structured estimation may be a more efficient option.

• Required precision: If high precision in work times is needed, time study or normal pace work sampling are the most suitable techniques. If a greater margin of error can be tolerated, structured estimation may be sufficient.

• Cost and time: Work sampling and time study require a greater investment of time and resources than structured estimation.

• Data availability: Structured estimation requires standard data or reference works to be effective.

The choice of the most suitable work measurement technique is a strategic decision that should be based on the specific needs of each company and the goals being pursued.