Nbcot Review Ch 2: Research

Front 1

What is an independent variable?

Back 1

- the presumed cause of the dependent variable
- the treatment you're using to make changes in the subjects/clients

Front 2

What is a dependent variable?

Back 2

- the phenomenon you seek to understand, explain, or predict.
- the assessments you are using to measure the effects of the treatment (ROM, MMT, other)

Front 3

Explain the difference between IV and DV.

Back 3

The dependent variable DEPENDS on the independent variable.
i.e. changes in ROM (dependent variable) due to treatment (independent variable) of PROM, AAROM, HEP

Front 4

List the types Quantitative Designs.

Back 4

1. True-experimental
2. Quasi-experimental
3. Non-experimental / coorelational

Front 5

Describe a true-experimental design.

Back 5

- Random selection and assignment into groups that either receive treatment or a control group that receives no treatment.
- The levels of treatment = IV
- The outcome that might be influenced by tx = DV

Front 6

Describe a Quasi-experimental design.

Back 6

- IV manipulated to determine its effect on DV, but there is a lesser degree of researcher control and/or no randomization.
- Used in health care when its unethical to control/withhold treatment.

Front 7

Describe a non-experimental/correlational design.

Back 7

- No manipulation of IV; randomization and researcher control not possible.
- To study potential relationships between two or more variables
- Cannot establish cause and effect
- Correlational coefficient (-1.00 to +1.00)

Front 8

What is a Qualitative Design?

Back 8

- A form of descriptive research that studies people, individually or collectively, in their natural social and cultural context.
- To describe real life experiences and give them meaning.
- Based on direct observation in naturalistic settings.

Front 9

List the types of Qualitative design.

Back 9

1. Phenomenological
2. Ethnographic
3. Heuristic
4. Case study

Front 10

Phenomenological qualitative research.

Back 10

A study of one or more persons and how they make sense of their experience
- meaning ascribed by participants only (not researcher)

Front 11

Ethnographic.

Back 11

Patterns and characteristics of a cultural group, including values, roles, beliefs, and normative practices are intensely studied.
- Extensive field observations, interviews, participant observation, examination of literature and materials, cultural immersion.
- Used in health care to understand an insider's perspective to develop meaningful services (study of a SNF)

Front 12

Heuristic.

Back 12

Complete involvement of the researcher in the experience of the subjects to understand and interpret a phenomenon.
- To understand human experience and its meaning.
- Idea that meaning can only be understood if personally experienced.

Front 13

Case study.

Back 13

A single subject or a group of subjects investigated in an in-depth manner.
- Purpose can be description, interpretation, or evaluation

Front 14

Describe statistics include:

Back 14

- Measures of central tendency
- Measures of variability

Front 15

What are measures of central tendency?
- Include the types.

Back 15

A determination of average or typical scores.
- Mean: the arithmetic average of all scores.
- Median: the midpoint of scores
- Mode: the most frequently occurring score

Front 16

What are measures of variability?
- Include the types.

Back 16

A determination of the spread of a group of scores.
- Range: different between highest and lowest score.
- SD: determination of variability of scores (difference) from the mean
- Normal distribution: symmetrical bell-shaped curve indicating the distribution of scores (mean, median, mode are all similar) *68% scores fall w/in +1 or -1 SD
- Percentiles: data divided in 100 equal parts and position of score determined.
- Quartiles: data divided into 4 equal parts and position of score determined.

Front 17

What are inferential statistics?
- List the types

Back 17

Determine how likely the results of the study of a sample can be generalized to the whole population.
- Standard error of measurement
- Tests of significance (alpha level, degrees of freedom, errors)
- Parametric statistics
- Nonparametric statistics
- Correlational statistics

Front 18

Standard error of measurement.

Back 18

An estimate of expected errors in an individual's score: a measure of response stability or reliability

Front 19

Tests of significance

Back 19

- An estimation of true differences, not due to chance.
- A rejection of the null hypothesis.

Front 20

Alpha level (p value).

Back 20

- Pre-selected level of statistical significance.
- What you as the researcher are willing to accept as the expected difference due to chance (the percentage of error)
- Usually .05 or .01
- .05 = only 5 times out of 100 or a 5% chance of error

Front 21

Degrees of freedom.

Back 21

Allows the determination of level of significance based on consulting appropriate tables for each statistical test.

Front 22

Errors.
- standard error
- type I
- type II

Back 22

- Standard error: the expected chance variation among the means; the result of a sampling error.
- Type I error: null hypothesis rejected when it's true (means of scores concluded to be different when they are not)
- Type II error: null hypothesis not rejected when it's false (means of score concluded to not be different when they actually are)

Front 23

What are Parametric Statistics?
- List the types

Back 23

Testing based on population parameters
- T Test: test of significance to compare two group means and identify a difference at a selected probability level (i.e. p=.05)
- ANOVA: compare two or more treatment groups or conditions at a selected probability level
- ANCOVA: compare two or more treatment groups or conditions while also controlling for the effects of intervening variables (i.e. two groups compared on UE reach with two types of assistive devices but one group has longer arms than other group so need to control for this)

Front 24

What are Non-Parametric Statistics?

Back 24

Testing not based on population parameters. Used when parametric assumptions cannot be met. Less powerful than parametric tests.
- Chi-Square Test: test of significance to compare data in the form of frequency counts occurring in two or more mutually exclusive categories (subjects rate treatment preferences)

Front 25

What are Correlational Statistics?
- Include testing measures, strength of relationships

Back 25

Used to determine relationships between two variables.
- Pearson used to correlate interval or ratio data
- Spearman used to correlate ordinal data
- Intraclass correlation coefficient (ICC): reliability coefficient based on an analysis of variance.

Strength of relationships:
- Positive correlation: 0 to +1, as x increases, so does y
- Negative correlation: -1 to 0, as x increases, y decreases
- High correlation: .70 to +1
- Moderate: .35 to .69
- Low: 0 to .34
***0 means NO RELATIONSHIP between variables

Front 26

What are the levels of measurement/data?

Back 26

Nominal data
Ordinal data
Interval data
Ratio data

Front 27

Nominal data.

Back 27

Classifying observations into mutually exclusive categories.
- measuring gender as female or male

Front 28

Ordinal data.

Back 28

Ranking. Numerical value that assigns an order to a set of observations.
i.e. Ranking income into categories (1=poor, 2=lower income, 3=middle income, 4=upper income). We can say that middle is higher than lower, but not the extent to which the rankings differ.

Front 29

Interval data.

Back 29

- Indicates how much categories differ. Uses equal spacing between categories.
- Absence of a true zero.
i.e. Fahrenheit and Celsius scales, IQ scales. There are no true zeros, but equal distance between categories.

Front 30

Ratio data.

Back 30

- The highest level of measurement. Has all the characteristics of the previous levels and a true zero.

Front 31

Normal Distribution.

Back 31

Draw Scale.

Front 32

Normal Distribution with averages.

Back 32

Draw Scale.

Front 33

To Test for Significance:

Back 33

- It doesn't matter what type of statistic we're calculating (T-test, chi-square, F-statistic, etc.), the produce is the same:
1. Decided on the alpha level you will use (error you are willing to accept)
2. Conduct the research
3. Calculate the statistic
4. Compare the statistic to a critical value obtained from a table

Front 34

If your statistic is HIGHER than the critical value from the table:

Back 34

- Finding is significant
- Reject the null hypothesis
- Probability is small that the difference/relationship happened by chance, and p is less than the alpha level (p<alpha)

Front 35

If your statistic is LOWER than the critical value from the table:

Back 35

- Finding NOT significant
- Fail to reject the null
- Probability is high that the difference/relationship happened by chance, and p is greater than the alpha level (p>alpha)