Stats 110 Exam 1 Flashcards

Population

The entire group about which the researcher wants information.

Sample

The individuals actually observed or providing usable data.

Sampling frame

The list or database from which the sample is selected. It may differ

from the target population.

parameter

A numerical description of a population; usually unknown. Common

notation: p^

statistic

A numerical description computed from a sample. Common notation: p^

individual

One person, object, animal, or other unit about which data are

collected.A characteristic recorded for each individual.

variable

A characteristic recorded for each individual.

Statistical Inference

Using sample information to draw conclusions about a population.

Descriptive Statistics

Organizes, displays, and summarizes observed data.

Inferential Statistics

Uses sample results to make conclusions or predictions about a population.

Categorical Variables

Places individuals into groups. Nominal (Groups data into names, labels, or qualities that cannot be sorted or ranked.) categories have no natural order; ordinal (Places data into categories that follow a natural sequence or hierarchy) categories do.

Quantitative Variables

Numerical values for which arithmetic is meaningful. Discrete values are countable; continuous values are measured on a continuum.

Observational Studies

Observes without imposing a treatment; cannot establish cause and effect definitively.

Experimental Studies

Deliberately imposes treatments and measures responses; a well-designed randomized experiment can support causal conclusions.

Census

A complete count or study that collects information from every single member of a an entire population.

Sample Survey

a method of collecting data from a sample to make valid inferences about a population.

Convenience Sampling

Definition: Selects individuals who are easiest to reach.

Clues and Concerns: Easy availability; usually biased

Voluntary Response Sampling

Definition: Individuals choose themselves whether to participate.

Clues and Concerns: Call-in or online poll; strong opinions are often overrepresented.

Simple Random Sample (SRS)

Definition: Every possible sample of size n has an equal chance of selection.

Clues: Lottery, random digits, or random-number generator.

Stratified Random Sample

Definition: Divide into homogeneous strata and take an SRS from every stratum

Clues: Ensures representation; compare subgroups. (doesn’t survey the collective group)

Cluster Sample

Definition: Divide into natural, internally diverse clusters; randomly select some clusters and survey all members in them.

Clues: Saves time and cost.

Systematic Sample

Definition: Choose a random start, then select every kth member.

Clue: “Every 25th name.”

SRS Random-Digit Table Steps

1. Label each individual in the population with a sequence of digits.

2. If there are 231 individuals in the population, label individuals with 3-digit sequences, such as 001 to 231.

3. Begin looking in the table at a random spot and read across by 3-digit sequences.

4. If the sequence you read corresponds to a labeled individual, select it for the sample.

5. If the sequence does not correspond to any label, or if the label has already been selected, ignore it and read on.

6. Stop when you have the needed number of individuals in the sample.

SRS Random-Digit Example:

Suppose you want to take a simple random sample of size 5 from 30 students in your Zumba exercise class. You label the students using two digits, 01 to 30, in alphabetical order. For uniformity, use line 101 of the table of random digits:

19, 22, 05, 13, 25

Random Sampling Error

Chance difference between a sample statistic and the population
parameter. It is reflected in the margin of error and generally
decreases with larger random samples.

Systematic Sampling Error

Persistent over- or underestimation caused by a biased selection
process, including undercoverage and poor sampling designs.

Nonsampling Error

Error not caused by random selection: nonresponse, inaccurate
responses, leading wording, interviewer influence, and data-entry or
processing mistakes. It can occur even in a census.

Bias

A method systematically favors certain outcomes. Larger samples do not repair a biased design.

Variability

Different random samples produce different statistics. Increase the
random sample size to reduce variability.

Margin of Error

a statistical measure that quantifies the maximum expected difference between a sample statistic (like a sample mean or sample proportion) and the true population parameter

What the margin of error does and does not measure.

The reported margin of error accounts only for random sampling variability. It does not account for undercoverage, nonresponse, voluntary response, misleading wording, interviewer effects, or
data-processing mistakes.

Calculating p^

Calculating Margin of Error (not 95%)

Quick Calculation for MOE 95% Confidence Level

Calculating Confidence Interval

How Sample Size and Confidence Level Affect Interval Width.

Larger Sample Sizes Make Confidence Intervals Narrower.

Response Variable

The measured outcome; also called the outcome or dependent variable.

Explanatory Variable

The factor thought to explain or cause changes; also called the
predictor or independent variable.

Randomization

Uses chance to assign treatments; helps balance other variables and reduces systematic differences

Control

Provides a baseline and holds other conditions as similar as possible.

Replication

Uses enough experimental units to reduce the influence of unusual
individuals and chance variation.

Completely Randomized

Assign all experimental units directly to treatments at random.

Randomized Block

First group similar units into blocks; then randomize treatments
separately within each block.

Matched Pairs

Compare two treatments using closely matched pairs or the same
subject under both treatments.

Placebo Effect

A response caused by the expectation of treatment rather than an
active ingredient.

Lurking Variable

An unmeasured or unrecognized variable that may influence the
observed relationship.

Confounding

Effects of two variables on the response cannot be separated.

(Researchers are aware of these variables)

Single-Blind

Participants do not know their assignments.

Double-Blind

Participants and the people interacting with them or assessing outcomes do not know assignments.

Triple-Blind

participants, researchers, and analysts are unaware of assignments

Statistical Significance

A result is statistically significant when an observed effect would rarely occur by chance alone under the relevant no-effect explanation.

Statistical significance does not guarantee practical importance, perfect methods, or replication of the result.

Preclinical Phase

In vitro and in vivo research before testing in humans.

Phase 1

First human testing; safety, toxicity, dosage range, side
effects.

Phase 2

Initial effectiveness, dose finding, continued safety and
tolerability.

Phase 3

Confirm effectiveness; compare with usual treatments;
monitor efficacy, toxicity, and side effects.

Phase 4

Post-marketing study of risks, benefits, adverse effects, and
optimal use.

Nuremberg Code (1947)

Fundamental ethical guidelines developed after Nazi medical
experiments.

Declaration of Helsinki (1964)

Strengthened ethical standards for medical research involving
humans.

FDA oversight (1960s–1970s)

Formalized drug approval processes requiring randomized controlled trials.

Belmont Report (1979)

Respect for persons, beneficence, and justice.

Institutional Review Board (IRB)

Reviews a study before it begins; evaluates risks, ethical treatment,
and informed consent.

Informed consent

Participants receive clear information about purpose, risks, benefits,
and the right to leave before agreeing.

Data Safety Monitoring Board

Independently monitors some ongoing trials and may recommend
early stopping for benefit, harm, or futility.