M4. L7. 

Types of sampling methods in Laboratory Scientific Research are primarily focused on controlling variables, ensuring the collected sample is representative, and achieving statistical validity. Unlike large-scale field studies, lab sampling often deals with small volumes, homogeneous solutions, or controlled populations.

SAMPLING TYPES

Random sampling in laboratory experiments is primarily used to ensure that samples selected for measurement are unbiased and representative of the entire population (e.g., a batch of solution, a population of cells, or a group of test subjects).

While laboratory settings often deal with smaller, more controlled populations than field research, the core subtypes of random sampling still apply to guarantee statistical rigor.

• Simple Random Sampling

This is the most basic form of random sampling, where every element in the population is given a unique identifier, and a random mechanism is used to select the sample units.

Scenario:

A researcher has 80 identical vials containing a chemical compound. They need to analyze 15 of them for quality control.

Procedure:

The researcher assigns each vial a number from 1 to 80. They then use a computer program or a table of random numbers to select 15 unique numbers, and the corresponding vials are tested.

• Stratified Random Sampling 

This method involves dividing the population into non-overlapping subgroups (called strata) that are homogenous (alike) with respect to a specific characteristic, and then taking a simple random sample from each stratum.

Scenario:

An experiment uses a new drug on a cell line that has been divided into two separate incubation chambers (Chamber A and Chamber B), which might have slight CO2​ or humidity differences.

Procedure:

The researcher randomly selects 10% of the cells from ChamberA and 10% of the cells from ChamberB. This ensures that any potential, though unknown, variations between the two chambers are represented equally in the final data set.

• Cluster Sampling

In laboratory research, this is less common than SRS or Stratified Sampling, but it is applied when the subjects or samples are naturally grouped or clustered for practical reasons.

Scenario:

A study is testing a new cleaning agent's efficacy on a large set of standardized bacterial samples grown on 50 different micro-titer plates (eachplate is a cluster).

Procedure:

The researcher randomly selects 10 of the 50 plates (clusters) to apply the cleaning agent to. All bacterial samples on those 10 plates are then measured for viability.

• Systematic Sampling

This involves selecting samples at a regular interval after a random starting point is chosen. It is often used for sampling over time or space in a continuous process.

Scenario:

A continuous flow reactor is synthesizing a product, and the quality needs to be checked throughout the 8-hour run. N=480 minutes, n=24 samples. k=480/24=20.

Procedure:

The researcher randomly chooses a starting minute between 1 and 20 (e.g., minute 7). They then collect a sample at minute 7, minute 27, minute 47, and so on, every 20 minutes until the run is complete.

SAMPLING METHODS

1. Serial Dilution

Serial dilution is a specific laboratory technique used to reduce the concentration of a sample in a controlled, step-wise manner. It isn't a sampling method in the sense of selecting individuals from a population (like random sampling), but rather a sample preparation method essential for accurate measurement when the original concentration is too high to count or read.

Its primary goal is to produce a measurable sample that yields results within the detection limits of the instrument or assay.

• Simple (Linear) Serial Dilution

This is the most common subtype, where the dilution factor is the same at each step. This creates an arithmetic progression in the denominator (e.g., 101,102,103).

Decimal Dilution (1:10 or 10−1)

A 1 part sample is added to 9 parts diluent (e.g., 1 mL of culture into 9 mL of water). Each step represents a 10-fold reduction.

Scenario:

A researcher needs to find the CFU/mL of a bacterial culture known to be millions of cells per mL.

Procedure:

Start with a 1:10 dilution. From that tube, take 1 mL and add it to a fresh 9 mL diluent tube (now 1:100). This continues until plates from dilutions like 10−5 or 10−6 yield a countable number of colonies (typically 30 to 300).

Two-Fold Dilution (1:2)

A 1 part sample is added to 1 part diluent. Each step halves the concentration.

Scenario:

Performing a minimum inhibitory concentration (MIC) assay for a new antibiotic against a pathogen.

Procedure:

Start with a 1:2 dilution. Subsequent steps are 1:4,1:8,1:16, etc. The final set of tubes contains a geometrically decreasing concentration of the drug, allowing the researcher to pinpoint the lowest concentration that inhibits growth.

• Geometric Serial Dilution 

This subtype uses a dilution factor that changes at each step, although this is less common for routine quantification and more for specific experimental designs.

Specific/Variable Dilution Factor

The dilution factor is calculated specifically for each step to target certain measurable concentrations based on a known non-linear dose-response curve.

Scenario:

A researcher is preparing standard solutions for a spectrophotometer where the analyte's absorbance is known to be non-linear at higher concentrations.

Procedure:

A stock solution might be diluted 1:2, then the next step 1:5, then 1:10. This focuses more of the measuring points in the lower, more linear range of the instrument's detection limit to improve the calibration curve's accuracy.

NOTES:
Why is Serial Dilution Necessary?

• Avoids "Too Numerous to Count" (TNTC): For microbial work, plating an undiluted sample results in a lawn of growth, making individual colonies impossible to count.

• Keeps Measurement in the Linear Range: Many instruments (like spectrophotometers) only provide an accurate, linear reading within a specific concentration range. Serial dilution ensures the measured solution falls within this range. Readings outside the linear range are often unreliable.

• Saves Material: Using a small aliquot from a highly concentrated sample and diluting it multiple times is more cost-effective and conservative of the original, often precious, sample.

2. Aliquot Withdrawal

Aliquot withdrawal is another common method, often used as a technique within a stratified or systematic sampling type. This involves the controlled removal of a small, precise volume, often using a calibrated pipette or syringe, from a larger volume like a reaction vessel or cell culture flask. For example, a researcher monitoring reaction kinetics might systematically withdraw a 100μL aliquot every five minutes from a well-stirred solution for immediate pH or OD measurement.

3. Direct Swabbing/ Plating

Direct Swabbing/Plating is a straightforward technique, where the method involves collecting a sample directly from a surface (swabbing) or transferring a liquid culture (plating) onto a medium. A subtype here is Spread Plating, where a sample (e.g., 100μL) is evenly distributed across a solid agar surface to grow isolated colonies.

4. Homogenization

Homogenization is a vital preparatory method that ensures the sample is uniform and representative before any measurement can take place. Subtypes include Mechanical Homogenization (e.g., using a tissue grinder or blender) and Chemical Lysis (using detergents to break cell walls). The procedure being to blend 1 gram of plant tissue in 10 mL of buffer before centrifugation to obtain a measurable extract.

Types of sampling methods in Engineering Prototype Research are fundamentally driven by the need to validate function, stress-test limits, and predict reliability of a newly designed solution. The focus is less on finding statistical relationships in nature and more on demonstrating that the prototype meets its specified design requirements under a variety of real-world or simulated conditions.

SAMPLING TYPES

1. Random Sample Testing: Selecting a random batch of prototypes from a production run to ensure manufacturing variability doesn't skew results (e.g., picking 10 chips from a batch of 100 for quality assurance).

2. Purposive/Critical Case Testing: Deliberately selecting units or scenarios that represent the most extreme, stressful, or required condition to validate critical functionality (e.g., testing the device only at its maximum rated temperature).

3. Consecutive/Time Sampling: Collecting data from the prototype over a continuous, extended period to monitor stability and drift, often used in reliability trials (e.g., logging sensor output every hour for a month).

4. User/Usability Sampling: Selecting a group of target users to interact with the prototype to assess human factors like ergonomics and ease-of-use (e.g., assigning 20 people to test a new control panel).

SAMPLING METHODS

1. Destructive Sampling (Testing)

This method involves testing a sample to its point of failure to determine its limits and structural properties. The sample is consumed or ruined during the test.

• Material/ Structural Validation

Determining the Tensile Strength (maximum pulling force) and Yield Strength of a 3D-printed plastic component.

Sampling Procedure:

A random batch of 10 prototype units is selected. Each unit is clamped into a Universal Testing Machine (UTM) and pulled until it breaks. The measurement is the force required to fracture the material (MPa).

• Stress Limit Testing

Measuring the maximum pressure a prototype fluid container can withstand before bursting.

Sampling Procedure:

A sample tank is filled with water and subjected to increasing hydrostatic pressure until the wall ruptures. The measurement is the Maximum Burst Pressure (kPa or MPa).

2. Non-Destructive Sampling (Testing)

This method assesses the integrity and quality of a prototype without causing damage or altering its function. The tested sample can often be used in the final product or for further testing.

• Quality Control/ Flaw Detection

Checking a welded joint on a prototype metal frame for internal defects like cracks or porosity.

Sampling Procedure:

The prototype is subjected to ultrasonictesting or X−rayradiography to image the internal structure of the weld. The measurement is a qualitative assessment of flaw presence/size, often against an industry standard.

• Dimensional Verification

Confirming that the physical dimensions of a newly machined part meet the Computer−Aided Design(CAD) specifications.

Sampling Procedure:

The sample is measured using a Coordinate Measuring Machine (CMM) or high-precision calipers. The measurement is a quantitative report of length, width, and tolerance deviation. 

3. Reliability/ Durability Sampling (Life Testing)

This involves subjecting the prototype to repeated use cycles or extended continuous operation to predict its lifespan and failure rate under normal or accelerated conditions.

• Endurance Testing

Determining how many open/close cycles a new hinge mechanism can tolerate before performance degrades.

Sampling Procedure:

A small, representative batch of hinges is installed in an automated testing rig. The machine continuously cycles the hinge (e.g., 100,000 times). The measurement is the number of cycles completed before failure, contributing to a Mean Time Between Failure(MTBF) prediction.

• Environmental Cycling

Testing the durability of a sensor designed for outdoor use under extreme temperatures.

Sampling Procedure:

The prototype is placed in a thermalchamber and rapidly cycled between high (50 °C) and low (-20 °C) temperatures for 100 hours. The measurement tracks the drift or failure of the sensor's readings after each cycle.

4. User/ Usability Sampling (Human-Centric)

This method involves collecting data from human users interacting with the prototype to identify flaws in the design's interface, ergonomics, or functionality.

• Ergonomic/ Interface Validation

Testing the intuitive nature and comfort of a new handheld consumer device.

Sampling Procedure:

A purposive sample of target users (e.g., 20 people of different ages and hand sizes) is asked to perform a set of tasks with the prototype. The measurement includes task completion time (seconds), error rate (%), and subjective user satisfaction scores (Likert Scale).

• A/B Testing (Iterative Sampling)

Comparing two slightly different versions of a software interface (A and B) during the prototyping phase.

Sampling Procedure:

Users are randomly assigned to use either Version A or Version B. The measurement tracks objective data like click-through rates and navigation paths to determine which design is more efficient.

Ethical considerations in Sampling and Measurement are absolutely vital, serving to protect the welfare of all subjects, ensure data integrity, and prevent the misrepresentation of research outcomes. Violations in these areas can severely undermine a study's validity and cause actual harm. The ethical principles apply across two distinct phases: sampling and measurement.

Ethical Considerations in Sampling (Subject Selection)

When selecting human or animal subjects, the process must be characterized by fairness, transparency, and respect:

1. Informed Consent and Voluntary Participation

Participants must be fully informed about the research purpose, procedures, risks, and benefits before consenting. Sampling methods must avoid coercing individuals, especially vulnerable populations, and researchers must obtain documented, informed consent.

2. Justice and Equity

The selection of subjects must be fair. The burdens of the research must not fall unfairly on one group while the benefits accrue to another. Purposive sampling must be scientifically justified to avoid demographic bias, ensuring that the sampling frame is inclusive and offers equal opportunity to all relevant populations.

3. Privacy and Confidentiality

Subjects' personal information must be rigorously protected. If data is collected from public sources, the sampling and measurement process must ensure the data is anonymized and de-identified before analysis. Coding systems should be used instead of names, and data must be stored securely to prevent breaches.

Ethical Considerations in Measurement and Data Integrity

These concerns focus on the honest and rigorous collection, recording, and reporting of quantitative data:

1. Avoidance of Fabrication, Falsification, and Plagiarism (FFP)

Researchers have an obligation to report data truthfully. This means never fabricating (making up) data or trials and never falsifying (manipulating or deliberately omitting) inconvenient measurement data, such as excluding an outlier just because it contradicts the hypothesis. All raw measurement logs must be preserved for audit.

2. Rigorous Method Application

The measurement protocol must be applied consistently and without bias across all samples and trials. The ethical obligation to achieve high Precision requires ensuring instruments are properly calibrated and that all systematic errors are eliminated or accounted for in the final reported results.

3. Transparent Reporting

All aspects of the research, especially its limitations, must be reported clearly for external evaluation. It is unethical to fail to transparently report the Sampling Method used, any Modifications to published measurement techniques, or any known bias in the sample or limitation of the measurement instrument.

Filling out a detailed table about Sampling Methods and Measurements is crucial for ensuring research is reproducible and validated. This table provides the operational blueprint of the experiment, documenting the exact how and when data was collected. This level of detail confirms that the measurements are accurate, unbiased, and compliant with international standards, such as the use of SI Units, which is key for scientific communication.

1. Parameter Measured

This requires specifying the exact characteristic being quantified (e.g., pH, velocity, plant height). The measurement must be precise; vague terms like "growth" should be replaced with exact metrics like Optical Density (OD   600) or Biomass (g/L).

2. Sampling Method

This column details the technique used to select and isolate the portion of the system for measurement, with the requirement to ensure representativeness or control bias. Refer to the lecture on types and methods of sampling. 

3. Frequency of Measurement

This specifies the exact timing to characterize the parameter accurately. Examples include:

• Continuous Monitoring

Real-Time Sampling. Data points are collected and analyzed instantly as the process occurs, often using a sensor. Example: An HPLC machine logging detector signal every second to resolve component peaks during chromatography.

High Frequency Logging. Data is collected at a very rapid rate (e.g., kilohertz or megahertz) to capture rapid transients. Example: An oscilloscope recording voltage spikes at 106 samples per second during an electrical discharge test.

• Periodic Intervals

Fixed Time Points. Measurements taken strictly according to a clock or calendar. Example: A pharmacokinetics study drawing blood samples from a patient at 0.5, 1, 2, 4, 8, and 12 hours post-drug administration.

Equidistant Spacing. Samples taken at mathematically uniform spatial or temporal intervals. Example: Collecting soil samples from a 10 meter by 10 meter grid at 1 meter intervals along the x- and y-axes.

• Event-Based

Threshold Triggered. Measurement is initiated only when a predefined condition or limit is met. Example: Sampling an industrial waste stream only when the effluent pH drops below 5.0 or rises above 9.0.

Process Step Completion. Measurement is tied to the successful conclusion of a stage in a procedure. Example: Analyzing the purity of an intermediate chemical immediately after the filtration step is finished, before moving to the final reaction.

• Single Measurement

End-Point Analysis. A final, definitive measurement taken when an experiment or process has officially concluded. Example: Measuring the final tensile strength of a material after it has been fully cured and cooled.

Initial State Assessment. A single measurement taken at the start to establish a baseline. Example: Measuring the initial dry weight of a plant seed lot before starting a germination study.

4. Conditions of Measurement

All critical factors that could affect the reading must be detailed, and conditions must be quantified wherever possible. Examples are temperature, pressure, humidity, light exposure, mixing/ agitation rate, pH/buffer type, wavelength, flow rate, atmosphere, concentration, et cetera.