602 Set 2

the class of individuals a researcher wants to generalize about.

E.g. the population of a presidential election = eligible voters in the U.S.

Usually a numeric fact about the population

E.g. # of male voters in the U.S., # of democrats etc…

• part of the population
• g. eligible voters in Hawai’i

• a number computed from a sample
• g. # of male voters in Hawaii

• When the sample represents the population
• 2nd is sample size

• Probability methods must be used to ensure sample representativeness
• Simple random sampling is the most basic probability method

drawing at random without replacement

everyone in the population has an equal chance to be sampled.

1. The interviewer has no discretion at all as to whom they survey 2. There is a definite sampling procedure, and it involves the planned use of chance

Any systematic error (eg surveying every third student passing the coffee stand)

No. Biases can be, sampling error cannot. Needs to be quantified

With a simple random sample, the expected value for a sample statistic = corresponding population parameter.

X and Y mutually cause each other

Experiments are a quantitative method to establish the effect of a presumed cause.

Randomization “equalizes” subjects between conditions by sorting different types of individuals into each condition with equal probability given large N. Thus randomization eliminates “unobserved heterogeneity” (= all individual differences not measures or measurable in a study).

Factors are independent variables. Conditions = levels of an IV, values that an IV takes.