• If a company hires a psychologist for a question such as “should flavour be congruent to
scent” or obtain/assess “flavour ratings of different potato chips” or other food products.
Then you must be very careful of your sampling procedures
o For example: Nestlé’s marketing strategy will be dependent on your finding
Sometimes sampling procedures are really important, other times, less
• The most effective way to obtain a representative sample is via random sampling. A
procedures whereby each person has equal chance of being selected as part of the
• Random sampling is likely to be representative in large samples
• **the definition of randomness pertains to how you pick or select subjects, not
how representative your sample turns out to be.
• Sometimes random sampling is used for ethical reasons, not just to obtain
o i.e. treatment outcome research: a small number of people can receive a
treatment that would be beneficial
• randomness is the most fair way
• most experiments involve comparing differences between one or more groups
o i.e. do males and females perform differently on a particular type of task
• Whenever we examine differences between groups on some dependent measures or
variables, we rarely find that the groups are identical.
o i.e. IQ test or mental rotation test. Big gender differences
• Must always ponder this in inferential statistics: if the groups do differ, is that due to
operation of chance factors or does it represent a true or real difference between groups.
• Chance factors dictate who will be in the sample. Must be mindful of this when collecting
• Random sampling will ensure the numbers are close, but not the exact same if the
sample is too small. So question is, how accurate are they? How close are they? Is the
error of esti