(Stats - 1): Introduction to Statistics and Research Design

Sample

A term used to describe a certain and specific population.

EX) U of W PSYC-2101 students

Descriptive Statistics

Used to organize and summarize data that came from a sample.

It communicates what the data “looks like” more clearly, and could be done using a visualization (graph) or a summary (measure of central tendency, such as mean)

Inferential Statistics

Answering the general research question.

Tells you if the data you collected is meaningful, and allows for some sort of conclusion to be made on how your data relates to the population.

Variables

Any sort of observation of a physical, attitudinal, or behavioral characteristic that can take on many values.

Some are concrete and easy to measure such as height, weight, BP, etc.

Some are abstract, relatable but no way to capture the data such as motivation, happiness, boredom, etc.

Operational Definitions

Specifies the procedure used to measure or manipulate variables, which is important when variables are abstract.

Discrete Observations

Taking on whole numbers, specific values with nothing in between. Such as seeing how much questions you got correct or incorrect on a test.

Continuous Observations

Has a full range of values with points in between integers (0.5) Such as how long it took you to complete a test.

Has a sequential order.

Categories are of the same size and intervals between categories are of the same size.

Has interval and ratio variables.

Nominal Variables

Has a qualitative, not a quantitative difference. Having different categories with no order between.

Such as arranging students into groups by their major.

Ordinal Variables

Has a directional relationship between categories.

This allows observations to be categorized in directions, but no magnitude.

EX) Rankings in a race. 1st, 2nd, 3rd, etc.

Interval Variables

Variables that have an arbitrary zero point, meaning that the zero still has a value. It’s higher than a negative but less than a positive.

EX) Temperature, the value of zero doesn’t mean that there is no temperature.

Ratio Variables

Variables that have an absolute zero.

EX) Money, $0 means you have no money,

Independent Variable (IV) or Predictor Variable

Variables that establish different conditions or groups that are being compared. Think of the term predictor, as this is the group that aids our prediction.

The values/conditions of the IV are referred to as levels.

Levels need a minimum of 2 and can either be established by manipulation or observation at different natural levels.

EX) Manipulation, decaffeinated vs caffeinated coffee. Observation, coffee drinkers vs coffee abstainers.

Dependent Variable (DV) or Outcome Variable

Seeing how much the thing that are we observing/measuring has changed/is different. Think outcome as in.. the outcome of the experiment.

Confounding Variables

Other variables that exist that could be changing at the same time as the IV, AND could affect the observed value of the DV.

Reliability

Receiving consistent, same results after measuring something multiple times.

Validity

Measurements that ensure that you are trying to measure what you NEED to be measured.

EX) Dog DNA kits have validity for determining breed, but do not have validity in determining personality.

What is the rule with validity and reliability?

A measure can be consistent (have good reliability), but it can also not have meaning (poor validity). Such as a scale that hasn’t been calibrated, giving the same wrong number twice.

However, a measure that has poor reliability (inconsistent) CANNOT have validity.

A good measure should be both reliable and valid.

Hypothesis Testing

The Research process involves:

1. Developing a Hypothesis

2. Defining the Variables

   1. What is the IV? DV?

3. Making observations of our variables

   1. Process should be reliable and valid

4. Hypothesis Testing

   1. Drawing conclusions about whether out data supports the hypothesis.

      1. Involves using a statistical approach that is more appropriate for that research strategy and design.

Experimental Research

Research involving the manipulation of an IV and observing the DV.

Correlational Research

Identifying the association between two or more existing variables. It also describes the existence using the direction and strength of the relationship.

Cannot infer causation, meaning that it cannot determine if one variable influences the other variable to happen.

Unethical/Impossible Research

Research that does not involve any manipulation. Such as testing hypothesis based on birth order.

How can we control confounding variables?

By the use of random assignment, basically assigning random individuals into random groups to hopefully minimize the impact of confounding variables, rather than sorting individuals based on certain experiences, traits, etc.

Between-Groups Design

Each individual is measured once during only one condition.

Within-Groups Design

Each individual is measured more than once, one for each condition. Also called repeated measures.

Data Ethics

A set of principles related to all stages of working with data such as design, collection, analyses, etc.

Open Science

An approach that encourages collaboration, including the sharing of research methodology, data, and stats in ways to allow questions from the public. It also includes preregistration (next slide.)

Preregistration

A requirement for clinical trials. Posting a published plan before the research starts.

What are unethical research practices?

• Conducting multiple analyses and choosing one that best matched the hypothesis

• Altering the hypothesis and presenting it as the original one (HARKing: Hypothesis After the Results are Known)

• Cherry picking data and excluding data that doesn’t match

• P hacking, which is like adding more individuals to a study even after testing has happened.