Research Methods Exam 2

Define IV, DV, & extraneous variables

IV – the variable that is manipulated, could influence or cause the outcome (DV)

 

DV – the variable (outcome) being measured

 

Extraneous – variables other than the IV that could affect the study’s outcome

Name 3 ways to control extraneous variables

1. Random Sampling – equal distribution of extraneous variables

2. Statistical Control (ANCOVA)

3. Blinding

single blind – the participants don’t know which group they are in

double blind – both the participants and the researchers don’t know which group the participants are in

Name & define the 2 general types of bias

1.  Researcher bias – intentional or unintentional – mislead the research either during data gathering or during data interpretation

2. Respondent bias– survey results are answered untruthfully

Define internal validity

The chance that observed changes are because of the intervention or treatment and not because of other possible causes

Define external validity

The chance that the results found in the subjects can be generalized to others who are similar

What’s the difference between inter-rater reliability & intra-rater reliability

Inter-rater reliability the extent to which different raters or observers perceive the same person or characteristic similarly

Intra-rater reliabilitythe degree to which each rater or observer is consistent in his or her ratings

Describe the 3 key elements of experimental design

1. Two groups, randomly assigned: experimental and control

2. Researcher manipulates IV in experimental group but not in the control group

3.Researcher measures DV to look for the impact that the IV had on the outcome

What’s the difference between double-blind & single-blind? Which is stronger?

single blind – the participants don’t know which group they are in

double blind – both the participants and the researchers don’t know which group the participants are in

 

Double blind is stronger

What’s the difference between systematic review & meta-analysis

Systematic Review – a study that summarizes and evaluates the existing research on a topic (Cochrane Database of Systematic Reviews)

Meta-analysis –similar to systematic review, but combines (pools) existing data and performs some new analysis.

What are quasi-experimental designs and why are they used?

One group pre-test post-test design – subjects serve as their own control: subjects received treatment (intervention) but also received a period of no treatment (intervention)

Cohort Designs – design is longitudinal; 2 naturally occurring groups (not randomized) followed over time; one group is experimental (intervention), one group is control

Time-Series/Repeated Measures –a “with-in subjects” design (one group);used to study the effects of treatment over time; no control group; subjects are tested under multiple conditions, and serve as their own control during periods of no treatment

name 2 quasi-exp designs that lack control group

One group pre-test post-test design and Time-Series/Repeated Measures

name a quasi-exp design that lacks randomization

cohort designs

Define cross-sectional longitudinal prospective & retrospective

Cross-sectional – a type of research in which you collect data from many different individuals at one point in time (survey)

Longitudinal – researchers repeatedly examine the same individuals to detect any changes that may occur over a period of time

Prospective – individuals are followed over time and data about them is collected as their characteristics or circumstances change

Retrospective-uses existing data that have been recorded for reasons other than research

Name 3 non-experimental designs.

Correlational research – relationship between 2 or more variables

Survey research – usually cross-sectional

Quantitative Case Study research –studies a person, group community or organization in depth and over time –longitudinal

Name & define the 4 types of data

Nominal – categories of data, with no order or relationship between categories (type of car, marital status, political affiliation, religion)

Ordinal – categories of data that can be ordered; however, the intervals between the categories are uneven or unknown (pain scales, satisfaction, manual muscle test)

Interval – ordered data with equal intervals between events; no true “zero”; data is usually continuous (temperature, time of day on a 12-hour clock)

Ratio –ordered data with equal intervals that also contains a zero point; possible to form ratios (weight)

Name the 3 conditions required to use parametric stats

Sample is representative and normally distributed

 

Variables generate Interval or Ratio data

 

Study groups are randomly assigned

What is the difference between population, target population, and accessible population?

Population – the larger group to which study results are generalized

Target population – represents the entire population for which any given study intends to examine

Accessible population –the group that the researcher can actually measure

What does it mean when a sample is considered representative?

Adequately represents the larger group according to whatever characteristic or quality is under study

Name two ways to increase the representativeness of a sample.

Inclusion Criteria – characteristics or traits that subjects must have to qualify for participation in the study

Exclusion criteria –undesirable features for participants

Name and define 4 probability sampling techniques.

Simple Random Sampling – all individuals who meet inclusion criteria are compiled and a sample of desired size is the selected at random

Systematic Sampling– use of sampling interval to select from randomly ordered possible subjects (total sample size needed divided by the available total population)

Stratified Random Sampling – population is first divided into subgroups (male/female/other) then each subgroup is randomly sampled, ensures representation from each subgroup

Cluster Sampling– used when you can’t identify individuals in a population; successive random sampling of groups or units (10 states → randomly choose 5 hospitals per state → randomly choose 5 therapists per hospital)

name & define 4 non-probability sampling techniques

Convenience Sampling – select cases based on their availability for the study and willingness to participate

Purposive Sampling  - Subjects are hand-picked based on particular attributes (often used in qualitative)

Snowball Sampling – Informants provide names of others who meet study criteria

Quota Sampling –divide population into subgroups based on characteristic(s); researcher sets quota for each subgroup (based on proportions in the population); then adds subjects until the quota is reached. (non-random)

why are descriptive stats used?

Used to describe, organize, and summarize data

 

Includes frequencies, percentages, measures of central tendency (mean, median, mode) and descriptions of relative position (range, standard deviation)

why are inferential stats used?

Used to make inferences (conclusions) about the population from the sample findings

Includes t tests, F tests, and tests for r

Describe symmetrical and asymmetric data distributions…what do they mean?

Symmetrical data – normal distribution (bell curve); mean, median and mode are similar

 

Positive asymmetrical data -  mean is to the right of the peak; tail is longer on the right; high number of low scores

 

Negative asymmetrical data –mean is to the left of the peak; tail is longer on the left; high number of higher scores

Define mean, median, mode, range, variance, and SD.

Mean – average of the scores

Median – midpoint of the distribution of scores

Mode –Most frequently occurring value

when are correlational stats used?

They are used to identify relationships (associations) between variables or sets of variables

Measure of association NOT causality!

explain strength & direction of correlation

Positive Relationship – value of r ranges from +0.01 to +1.00

As one variable increases, the other variable also increases.

Strength determined by closeness to +1.00

 

Negative Relationship – value of r ranges from -0.01 to -1.00

As one variable increases, the other variable decreases

Strength determined by closeness to -1.00

what is linerar regression? why’s it useful?

Used after a correlation coefficient has been established

If there is a relationship between variables, future scores of DV can be predicted

Define chi-square, t-test, & ANOVA

Chi-square -  used for non-parametric (nominal) data; tests of goodness of fit: compare observed categorical data (frequency) to an expected distribution… see if there is a difference

 

t-Test – used for parametric data; used to compare the mean scores of two groups

 

ANOVA –compares the mean scores of three or more groups

what are the 3 types of t-test and how are they used?

Single sample t-Test – compares sample mean to a known mean (such as a norm for a population)

 

Paired samples t-Test – subjects are compared to themselves (such as one group pre-test post-test)

 

Independent samples t-Test –two different groups are compared (such as experimental and control)

Name the non-parametric equivalents of 1) Pearson r, 2) unpaired t-test, 3) ANOVA

Pearson r – Spearman’s rho (r_s)

Unpaired t-Test – Mann-Whitney test (U)

ANOVA - Kruskal-Wallis test (H)

Scenario: A researcher designs an RCT (N=100) to see whether virtual reality is more effective than standard care for managing pain in pts with burns.

1. What type of design is this? 

2. What is the IV? 

3. What is the DV?

4. What is the research hypothesis?

5. What is the null hypothesis?

1.) Experimental Design

2.) Virtual Reality

3.) Pain in patients with burns

4.) That virtual reality is more effective than standard care for managing pain in patients with burns

5.) There is no difference between virtual reality treatment and standard care for managing pain patients with burns

The researcher collects data from both groups before and after an OT session, using a standardized  pain assessment.

1. Which stat could be used to compare means within groups?   

2. Which stat could be used to compare means between groups?

3. Did you choose parametric or non-parametric stats, and why? 

1.) Paired Samples t-test

2.) Independent Samples t-test

3.) Parametric – parametric t-tests are used to compare the mean scores of two groups

The researcher sets the level of significance (alpha) at .05 before data analysis begins. The calculated p-value for the between-groups statistical analysis  is p = .02

1. Should the researcher reject or not reject the null hypothesis? Explain how you decided.

2. What does this result mean, in your own words?

1.) Reject the null hypothesis because 0.02 is less than 0.05

2.) That there is a statistically significant difference between virtual reality and standard care in managing pain in a burn patient.