All true hypotheses

are falsifiable (specifying what a null result would look like, as well as your alternative)

Good Hypotheses

Are parisomious: only including variables, measures, and constructs that are evidenced, and add explanatory value Operationalize Construsts Well: actually get at the thing you are trying to measure Communicate an Idea well: getting a clear, singular idea across than a limit a methods

Post-hoc hypotheses

- developed after design/observation rather than before -potential statistical and conceptual issues -desirable when: exploratory, any time they are clearly stated as such, and are necessary to get at an important point

a-priori hypotheses

- at least specified (if not preregistered) before data collection has occurred -usually required for research (if you have an idea of what you're looking for, you must state it)

Exploratory Research

-sometimes an area of study or a particular question is too new to use existing findings to generate hypotheses -usually identifies constructs of interest, but may leave flexibility in how they are measured and analyzed

Research findings ARE:

-observational -specific and concrete -may be qualitative, quantitative, or non-experimental -may involve analysis

Research findings are NOT:

-assumptions or inferences -a proposed relationship between concepts -the same thing as a hypothesis that has been supported

Reading and Writing Findings

-usually presented as a result rather than in a discussion -regard specific use-cases of variables and measures -regard hypotheses, research questions, or exploratory analyses -may be the result of observations collected in one study, or may be formed through observations across studies

When presenting a finding:

-explain the empirical way the finding was supported -assess whether the finding explains a main point or hypothesis on supports/refutes one -identify any independent/predictor variables, dependent/outcome variables, or control variables that the finding regards clearly, properly cite the source of the finding

Identifying Findings in Articles

-found as results of analyses -usually can find a conceptual summary of what statistic was meant to assess and what it's result indicated before or after each statistic in a results section -will address each formally presented hypothesis through at least one finding

Findings and Causal Inference

-experimental approach -->manipulation of "variable x" observably and reliably impacts measurement of "variable y" -theoretical approach --> observational frameworks, explanatory and evidenced theory, elimination of possibly confounds and "third-variables"

Null Distribution

-the sampling distribution we construct for possible data under the null

Sampling Distribution

a graph of the frequency of possible observations for a given sample size, given infinite samples

Alternative Distribution

the sampling distribution of all possible observations not under the null

To find a p-value

take our actual observed data, and find it's position in the null distribution - the proportion of the area under the curve at or past this value is the same as the likelihood of drawing a value like this, or more extreme (unlikely) from the null distribution -we set a cutoff point at which a p-value is statistically significant (alpha level)

Type I error

incorrectly rejecting the null (when p <.05 but the null is correct)

Type II error

Failing to correctly reject the null (when p>.05 but the null is incorrect)

Error in setting alpha level

-if we set it too low, we may miss a correct alternative and making a type II (false negative) -if we set it too high, we may unduly reject the null and make a type I error (false positive)

Non-Experimental Research

the researcher observes and measures variables as they naturally occur without manipulating them or assigning participants to treatment groups

Observational Research

-describe a variable or set of variables -obtain a snapshot of specific characteristics of an individual, group, or setting

Naturalistic Observation

observing people in the environment in which it naturally occurs

Naturalistic/Participant Observation

Researchers become active members of the groups they are studying Ex: undercover bosses

Structured Observation

An investigator makes careful observations of one or more specific behaviors in a particular setting that is more structured than the settings used in naturalistic or participant observation

Case Study

-Can be an individual, social unit (cults), events (disasters) -good on short term memory, non long-term encoding -can provide insight, but are not regarded as a form of hypothesis testing

Archival Research

Analyzing data that has already been collected for some other purpose - census -open-source data (TVMI veteran initiative) -long- term longitudinal studies

Correlational Research

-science used to describe and predict -describe strength and direction between variables -usually refers to research where the IV cannot be manipulated -used to develop tools and make predictions

Hypothesis Testing with Pearson's r

1: establish an appropriate hypothesis 2: construct the null distribution 3: calculate the r-coefficient for two variables to test your hypothesis 4: use software to calculate the odds of obtaining an r-coefficient as extreme or more extreme than the value you obtained from the null distribution 5: reject or fail to reject the null, using your a-level as your criteria

Correlational Research in Context

-used to validate measures -can have much higher external validity -can provide converging evidence to support a theory

Correlation and Causation

1. Directionality Problem 2. Third Variable Problem -correlation doesn't equate to causation -correlation can suggest, or even test a causal relationship

Correlation Matrix

Can be used to check all variables in a study

Reliability Measures

Test intercorrelation among items, subsections, inter-rators, instruments, or studies. e.g. Chronbach’s-alpha

Factor Analysis

organizes the variables into a smaller number of clusters, such that they are strongly correlated within each cluster but weakly correlated between clusters. Each cluster is then interpreted as multiple measures of the same underlying construct. These underlying constructs are also called “factors.” Condensing variables into factors

Exploratory Factor Analysis

-forms clusters of variables or items in a measure -relies on researchers to decide how many of these clusters are meaningful, and what they indicate for exploratory analyses

Confirmatory Factor Analysis

-describes how well the cluster or factors accounted for a data set -these often (and sometimes problematically) exist in the same papers as the original exploratory analysis

Simple Regression

A statistical technique that allows researchers to predict the value of one variable given another

Directional Prediction

sed for predicting some “Y” from some number of “X”s.

Multiple Regression

Used to tell if predictor variable affects the outcome variable over and above other variables

Covariates in Research

-Each predictor variable in a regression accounts for some amount of the variance in the outcome variable -variables that researchers don’t want to test, but do want to account for as a way of eliminating confounds - most commonly used in correlational research using regression methods -can be used alongside an IV you can manipulate in an experiment

Qualitative Research

1. In-depth information about relatively few people 2. Conclusions are based on interpretations drawn by the investigator 3. Global and exploratory

Quantitative Research

1. Less depth information with larger samples 2. Conclusions are based on statistical analysis 3. Specific and focused

Triangulation

Both data types (quant/qual) should converge at some point