Sensory Analysis Flashcards
Sensory analysis
A scientific method used to evoke, measure, analyze, and interpret responses to products as perceived through the senses
Senses
Vision
Gustation
Olfaction
Touch
Audition
Gustation
Flavor = taste + aroma
The sense of taste involves perception of non volatiles
Taste receptors are on the tongue and mouth
5 Tastes
Salty
Sweet
Sour
Bitter
Umami
Salty
NaCl
KCl
Sweet
Sucrose
Glucose
Sweetener replacers
Sour
Acid
(Citric, phosphoric)
Bitter
Quinine
Caffeine
Umami
MSG
Potentiators
Enhance taste sensations
Umami (meat)
Salt (sweetness)
Acids
Attenuators
Inhibit perception
Sugar (acidity)
Fat (saltiness)
Miraculin
Sugar substitute
Glycoprotein
Makes sour food taste sweet
Binds to sweet preceptors in a sour environment
Olfaction
Volatile molecules sensed by receptors
Orthonasal olfaction
Breathing
Sniffing
Through the nose
Retronasal olfaction
Via the back of the throat
Volatiles go through mouth and into nose
Touch
Evaluate the consistency, texture, viscosity of foods
Audition
Noise emitted by food
Contributes to perceived texture
Expectation error
People tend to find what they expect to find
Give as little info as possible
Suggestion effect
Comments or noises made out loud affect judgments
Biases perception
Want physical separation
Distraction error
Conversations
Time pressure
Personal preoccupations
Be mindful of sample # and time of day
Color + intensity
Products of deeper color are presumed to be more intense in flavor
Use red light to prevent them from seeing the true color
Habituation
Need to give people breaks
Vary products
Order effect
Randomize and balance the order of presentation
Lingering aromas transfer between samples
Central tendency
Panelists avoid the scale extremes
Train the assessors
Other sources of error
Motivation error
Brand names can bias assessors
Rating of one attribute can influence rating of others
Adaptation
Limit the number of samples presented
Allow the sensory system to recover
Palate cleansers
Coding
Random 3 digit codes
Different codes for replicates of the same sample
Consistent format
Positioned in the same location
Palette cleansers
Avoid carry over
Bottled water
Milk
Apples
Saltines
Parts of testing facility
Sample prep area
Serving area
Booths for panelists
Discussion/training area
Storage area
Equipment
Sample temp
Hot foods
60-66 °C
Hot tea
66-71 ° C
Cold beverages
5-9 °C
`
Frozen desserts
-18 to -10 °C
Sample size
It’s not lunch!!
You need enough sample to measure the attributes
Things to keep in mind about panelists
Recruitment
Confidentiality
Untrained vs trained
Sample demographics
Types of data
Nominal (labels)
Continuous (any possible #)
Ordinal (ranking scale)
Data handling
Check raw data for errors
Outliers and missing values
Data transformation
Objective tests
Numerical data
Discrimination tests
Descriptive analysis
Subjective tests
Preferences, comments, reactions
Affective tests
Discrimination tests
Determine whether there are any sensory differences between samples
Discrimination tests used for
Screening and training assessors
Investinging taints
Determine sensitivity thresholds
Quality control (consistency of materials)
Preliminary assessment
Overall difference tests
Assessors use all available info to make decision
Can be restricted to one specific characteristic
Detect differences between samples
Triangle tests
Determine if a difference exists between 2 samples
Panelists presented with 3 samples (2 are the same)
Identify which sample is the odd one
24-30 assessors
Triangle test data analysis
Total # of responses correctly identifying the odd sample
Number of correct responses compared to statistical tables
Number of correct responses must > critical min value
Triangle test conclusion
A significant differences does or does not exist
State significance level
Duo trio tests
Determine if a difference exists between 2 samples
Panelists get 3 samples
1 is the reference sample
Which one is the most similar/different to the control?
Min of 32 assessors
Duo trio tests data analysis
Total # of correct responses counted
Number of correct responses compared to statistical tables
Must exceed critical value to claim a difference
Difference from control test
Determine if a difference exists between 1 + samples and a control
Determine the magnitude of difference between the samples and control
20-25 panelists (fewer if highly trained)
Difference from control data analysis
Mean score calculated for each sample
Difference in scores represents heterogeneity
Two factor anova
Same different test
Presented a pair of samples
Determine if the samples are the same or different
30-50 assessors
Attribute specific tests
Focus on one attribute or quality
A not A test
Two samples are presented
Control / not control
10-50 panelists, trained
2 AFC
Determine if a difference exists between 2 samples for 1 specific attribute
Present 2 coded samples
Identify which of the samples has a greater intensity of xyz attribute
Min of 30 panelists
2 AFC data analysis
Determine the total # of times each sample is selected
Larger # of responses compared to statistical table
Min # of response to conclude a statistical difference
2 AFC conclusion
One sample is more intense than the other or that there is no difference
Directional difference tests
Determine in which way a particular sensory attribute differs between 2 samples
3 AFC test
Determine if a difference exists between 3 samples with regards to a specific attribute
2 samples are the same, 1 is different
Ranking test
Determine if a difference exists between 3+ samples with regards to a specific attribute
Forced to make a choice for each position
Ranking test samples
Number of samples depends on how fatiguing the assessments are
order of sample prep should be balanced
Ranking test data analysis
Data are summarized in a table showing rank order
Rank orders summed and divided by # of samples tied for that position
Friedman statistic calculated
Descriptive analysis
Identify the nature of a sensory difference and/or the magnitude of the difference
Descriptive analysis key steps
Select and train assessors
Generate attributes/references
Agree on attributes
Determine assessment protocol
Rating intensity and scale design
Flavor profiling
Assess aroma, mouthfeel, flavor
4-6 trained panelists
5 point scale
Eval by yourself, discuss after, determine a consensus score
No stats!
Texture profiling
Texture and mouthfeel assessed
13 point scale
6-10 panelists
work in consensus
Quantitative descriptive analysis
Analyzed statistically
Ful quantitative and qualitative sensory description
8-15 trained panelists
Spectrum method
Extension to products outside of food/beverages
Full quantitative and qualitative sensory description
Sensory qualities are assessed using predefined and standardized lexicon
15 point scale
12-15 selected panelists
Displaying sensory data
Spider plots
Sensory traces
PCA
Spider plots
Each attribute represented
Center of plot = 0 perceived intensity
Attribute means are plotted and joined with continuous lines
Sensory traces
Attributes marked along X axis
Y axis = perceived intensity
Means are plotted and joined using a line
Affective tests
Consumer testing assesses subjective responses to a product
Quantitative affective tests
100 panelists
Usage/non usage
questionnaire or face to face
Focus group
Formulate a hypothesis
Test the feasibility of a new products
Identify attitudes, opinions, preferences
8-12 participants
Focus group procedure
Trained moderator produces and guides the discussion
Written report
Video or audio recording
Preference tests
Preference tests provide evidence of whether one product is preferred over another
50-100 panelists
2 products (paired test)
2 or more products (ranking test)
Hedonic ranking
Subjects asked to rate liking on hedonic scale
Responses converted to numeric values before analysis
100 panelists
Attribute diagnostics
Why do consumers like/dislike products
Which sample is preferred in terms of ___
How much do you like the products
