3: External Validity

Question 1

Population validity

Answer 1

Do the research findings obtained from a sample apply to the population of interest

Question 2

Ecological validity

Answer 2

Do the research findings reflect what people do in the real world
-influenced by subject reactivity, research setting or how variables are operationalized

Question 3

Population of interest

Answer 3

All individuals to whom we desire to generalize the findings of a research study
(Census: includes every member of a population)

Question 4

Sample

Answer 4

Smaller subgroup of subjects chosen from the population

Question 5

Representative sample

Answer 5

Closely match various chanarterisitcs of the population of interest

• allows to generalize to population from which drawn
• important for characteristics that affect studied variables

Question 6

Biased

Answer 6

When sample characteristics don’t match those of population

• stems from nonrandom procedures
• attempt to generalize outside a biased sample may be misleading or inaccurate

Question 7

Steps in sample selection

Answer 7

1. Recruitment: risk for sample bias
   • only part of target population will be accessible (sampling frame)
   • may not reflect POI
   • reach potential participants in some way (contacted sample) so aware of possibility
2. Enrollment: risk for non-response or volunteer bias
   • inclusion/ exclusion criteria
   • eligible part. Choose to part. Or decline

Question 8

Sampling bias

Answer 8

If contacted sample is not representative of target population

Question 9

Volunteer bias

Answer 9

Is participants are systematically different from non-participants

Question 10

Random sampling

Answer 10

Every member of POI has equal chance of chosen
-generates a representative sample, so generalized to population drawn

• requires a directory of sampling frame (delivery sequence file, phone book, list of area codes)
• assign numbers to each record
• then use probability sampling, and a random number generator

Question 11

Nonrandom sampling

Answer 11

Every member of POI interest does not have same chance of being chosen
-will be biased or unrep. So must be cautious when generazlizing

Question 12

Probability sampling

Answer 12

Use at least one element of random sampling to ensure representative
-crucial for frequency claims, external v important

Question 13

Simple random sampling

Answer 13

Randomly sample subset of individuals from population
-reduces sampling bias, but guarantee representative sample
• some segments over or under-represented by chance or method of recruitment

Question 14

Stratified random sampling

Answer 14

Population first divided into demo. Strata
• random sample of proportionate size Ns is drawn from each startum
(Each Ns is proportional to percentage of POI in each strata)

•ensures sample representative of demo strata even if diff sizes

Question 15

Oversampling

Answer 15

•population first divided into demo. Strata.
•a random sample of some fixed n is drawn from each
-smaller strata will be over-represented
- results later weighed according to proportion
•ensures small strata are sampled adequately

Question 16

Systematic sampling

Answer 16

From a directory of POI, randomly choose a starting point and interval k
• every Kth element after starting point is sampled

Question 17

Cluster sampling

Answer 17

• units of sampling are groups rather than individuals
• clusters randomly sampled from all clusters in population
• all individuals within chosen cluster are included
• useful when v large populations

Question 18

Multistage cluster

Answer 18

• cluster sampling followed by another sampling technique
  1. Randomly choose large clusters
  2. Randomly sample individuals from within those randomly chosen clusters (using simple,stratified, oversampling, systematic)

Question 19

Convience sampling

Answer 19

Sample is drawn from a more restricted sub population that is easy to access
-in sampling bias if convience sample is systematically differnt from POI

-findings may not generalize to other types of people or other types of students

Question 20

Self selection

Answer 20

Enrolling only those who invite themselves

-results in non-response or volunteer bias if diff from POI

Question 21

Biased sampling techniques

Answer 21

-conscience sampling
-relying on self-selection
-recruiting materials don’t cover population
-other bias sampling (not random):
Purposive: nonrandomly target certain kinds of ppl
Snowball: ask enrolled part. To recruits others in their social networks
Quota: nonrandom within each stratum until reaching target number of that stratum

Question 22

Lab studies

Answer 22

• generally afford greatest control over EV (higher degree of internal validity)
• artificial environment (questionable ecological)

Question 23

Field studies

Answer 23

-conducted outside lab, in real world
(High degree of ecological validity)
-difficult to control EV in field (difficult internal validity)

Question 24

Naturalistic observations

Answer 24

Observe and record naturally occurring behavior in field (no manipulation)
-not used to support causal claims low internal validity

Question 25

Field experiment

Answer 25

Test for casual relationships outside lab by manipulating variables in field

• attempt to control EV to degree possible
• diffciult to control EV, threatening internal and causal

Question 26

Lab simulations

Answer 26

Attempt to recreate the real world setting in lab
-able to study behavior under more realistic conditions (increasing ecological) mailing control EV
(Greater realism increase generalizability)
-lab simulations used instead of a field study when:
•internal is essential but difficult to ensure in field
•behavior cannot be ethically studied
•under naturally occurring conditions is expensive and time consuming