Research Flashcards Preview

NCE 2018 > Research > Flashcards

Flashcards in Research Deck (31)
Loading flashcards...
1
Q

Step 1:

The process of reporting research begins with a purpose statement that identifies the variables being studied and the population to which they pertain.
There are three common types of variables in social research:

A
  1. Independent – Independent variables are those that are manipulated or selected by the researcher to cause, influence, or otherwise effect the outcome.
     To identify the independent variable(s), ask, “What is the name of the theory or technique the researcher is using to cause change?”
  2. Dependent – Dependent variables are those that are affected or changed as a result of the manipulation of the independent variable(s).
     To determine the dependent variable, ask, “What is the researcher attempting to measure or test?”
  3. Control – Control variables are possible confounding variables that the researcher attempts to hold constant so that their effects are canceled out or controlled for, such
    as demographic or background characteristics of the subjects of a study.
     To determine the control variable(s), ask, “What is the demographic or background information identified in the purpose statement?”
2
Q

Step 2
The second step is to state the hypothesis. Pure experimental research is based on a null hypothesis — a statement that there is no true relationship existing between the independent and dependent variables.

A

Sample Null Hypothesis #1:
There is no significant difference between behavior therapy and physical therapy in reducing self-reported chronic lower back pain (CLBP), among adult males ages 40-60.
or
There is no significant difference in the reduction of self-reported chronic lower back pain (CLBP), among adult males ages 40-60, whether behavior therapy or physical therapy is used.

Sample Null Hypothesis #2:
There is no significant difference found between subjects’ scores on the communication scale with respect to their success in forming friendships as measured by a sociogram.

3
Q

Four Types of Sampling Techniques

A
  1. Simple Random Sample – In a Simple Random Sample each item/subject in a sample is considered to have an equal, independent chance of being selected “into”
    the sample.
  2. Stratified Random Sample – In a Stratified Random Sample representative items/subjects are divided into parts (grades, ages, income, test scores, etc.). In
    each part, each item/subject has an equal, independent chance of being selected “into” the sample.
  3. Cluster Sample – In a Cluster Sample parts that go together are researched/studied together (neighborhood, class, etc.).
  4. Systematic Random Sample – In a Systematic Random Sample a systematic rule of selection or predictable interval is employed (every 3rd person, odd numbered,
    etc.).
4
Q

Sampling Error and Sampling Bias

A

1a. Sampling Error – occurs when subjects are not under the researcher’s control or when a discrepancy arises due to random sampling.
1b. Sampling Bias – is considered to be the researcher’s fault when it occurs because it involves a researcher’s selecting a non-representative sample for his/her own convenience or due to his/her own prejudices.

5
Q

Threats to internal validity concern flaws in the design of the study. The following are types of threats to the internal validity of a study:

A
History
Maturation
Testing
Statistical Regression
Subject Attrition
6
Q

Threats to external validity concern the extent to which the researcher can generalize findings to a larger population. Common threats to the external validity of a study include the following:

A

Hawthorne (placebo) Effect – occurs when the subjects’ knowledge that they are participants in a study alters or otherwise influences their usual responses.
Control: Have some sort of irrelevant treatment for the control group.

7
Q

Threats to external validity concern the extent to which the researcher can generalize findings to a larger population. Common threats to the external validity of a study include the following:

A

Experimenter or Rosenthal Effect – occurs when the researcher/experimenter’s behavior or appearance affects the subject’s performance.
Control: Use more than one experimenter for interrater reliability.

8
Q

Threats to external validity concern the extent to which the researcher can generalize findings to a larger population. Common threats to the external validity of a study include the following:

A

Halo Effect – occurs when the researcher/ experimenter allows his/her initial impressions of subjects to influence later ratings of subjects.
Control: The researcher recognizes and notes the bias

9
Q

Types of Measurement Scales

A
A. Nominal or Categorical – is the lowest and least precise level of measurement as it simply classifies or sorts persons and/or objects into categories — names of categories — hence the term "nominal."
Examples: yes/no
male/female
married/never married
married/divorced/widowed
Group 1/Group 2/Group 3
B. Ordinal (Order) – not only classifies subjects or their behaviors, but also ranks them in terms of the degree to which they possess a characteristic of interest.
Intervals between ranks are not equal.
Examples: BA/MA/Ph.D.
1st Place/2nd Place/3rd Place
Freshman/Sophomore/Junior/Senior
Chapter 1/Chapter 2/Chapter 3/Chapter 4

C. Interval (Equal) – has all the characteristics of both a nominal and ordinal scale, but, in addition, it is based upon predetermined equal intervals.
Examples: 1, 2, 3 or 3, 2, 1
-1, -2, -3 or -3, -2, -1
20 degrees, 40 degrees, 60 degrees;
-20 degrees, -40 degrees, -60 degrees
Note: Most of the tests used in educational research, such as achievement tests,
aptitude tests, and intelligence tests, represent interval scales. Interval
scales do not have a true zero point (which means they CAN HAVE
negative numbers).

D. Ratio – is the highest, most precise, level of measurement. A ratio scale has all of the advantages of the other types of scales and, in addition, it has a true zero point
(NO negative numbers).
Examples: 5 lbs., 10 lbs., 15 lbs.
10 inches, 12 inches, 14 inches
25 cents, 50 cents, 75 cents, $1.00
10
Q

AB Design

A

is the simplest single-subject design. It includes one baseline phase (A) and one intervention phase (B).

11
Q

ABAB Design

A

is designed to better control for outside events. The 2nd “A” stands for a second baseline phase (which would occur after some period of time AFTER the withdrawal or removal of the intervention). The 2nd “B” stands for the RE-introducing of the intervention. So, if you were to be asked to interpret what “ABA” represents, you would deduce that there has been an initial baseline established, an intervention has been introduced, and that the researcher has now withdrawn the intervention. After some period of time, another baseline will be established - thus the second “A.”

12
Q

Quasi-experimental Research

A

Quasi-experimental Research – is widely used in the counseling field. Only the first of the two criteria listed above for experimental research is met since subjects are
usually not randomly assigned because the groups under study are already intact (classroom groups, etc.).

13
Q

Types of Derived Scores

Grade Equivalent

A

denotes that average raw scores are assigned a grade-level value. We interpret a given score compared to other “x” graders: This subject performed at an average/below average/above average level.

Example: So, if a third (3rd) grader, Mary, scores a 4.5 on a reading test, theappropriate way to interpret/report the score is to say, “Mary scored above average in reading.”

14
Q

Types of Derived Scores

Percentile Rank

A

indicates the percentage of scores that fall at or below a
given score.

Example: So, if John scored a 59 on a reading test and that score landed him at the PR of 45 (based on a graphing of the scores), the meaning is that 45% of the students who took the test earned scores of 59 or less.

15
Q

Types of Derived Scores

Standard Scores

A
  1. z-score – is the most basic standard score and allows scores from different tests to be compared. The z-score has a mean of zero and a standard deviation (SD) of
    one.
  2. T-score – is widely used and has a mean of 50 and a standard deviation (SD) of 10.
  3. Stanines – (contraction of the two words “standard nine”)
    divide the normal curve into nine parts,
    NOT NINE EQUAL PARTS.
16
Q

Measure of Central Tendency

A

Mean, Median and Mode

17
Q

Mean

A
Mean is the arithmetic average of the scores and is the most frequently used measure of central tendency because it is more stable and precise. The mean is used with INTERVAL and RATIO scales and is determined by calculation: Add or sum up the scores and divide
that number (the sum) by the number of scores.
18
Q

Mode

A

Mode is the score attained by more subjects than any other score. The mode is not established through calculation; instead, it is determined by looking at a set of scores or at a graph of scores and seeing which score occurs most frequently. A set of scores may have two (or more) modes, in which case it is referred to as bimodal or
multi-modal. Although the calculation of the mode can be done with any level of data, it is the only appropriate measure of central tendency that is used with NOMINAL data.

19
Q

Median

A

Median is that point in a distribution above which and below which are 50% of the scores; in other words, the median is the midpoint (middle). It is not frequently used,
but it is the best to use when a distribution has one or more extremely high or extremely low scores. The median is the most appropriate measure of central tendency for ORDINAL level data.

20
Q

Standard deviation

A

Standard Deviation is the most often used measure of variability as it describes how scores vary around the mean. The larger the SD, the greater the variability among
scores.

SD is a measure of Variability

21
Q

Variance

A

Variance is one way to measure or describe the variation of/between data (just like the term standard deviation describes how data vary around the mean/median/mode).

22
Q

Measurement of relationship

Postivie and negative

A

If two variables are positively and directly related (as X increases, Y also increases or as X decreases, Y also decreases), the correlation coefficient will be close to +1.0, a perfect positive relationship.
Example: The relationship between exercise and muscle tone.

If two variables are negatively and directly related (as X increases, Y decreases, or as X decreases, Y increases), the correlation coefficient will be close to -1.0, a perfect negative or inverse relationship.
Example: The relationship between amount of exercise and percentage of body fat.

23
Q

Measurement of relationship

These are the two widely used correlations:

A

a. The Pearson product-moment correlation (Pearson r), which is used for interval or
ratio measures.
b. The Spearman rho, which is used for ordinal data.

24
Q

Skewed distributions

Negative and positive skew

A

Negative Skewed:
The mean is pulled in the direction of low scores (the tail to the left). If a distribution is negatively skewed, the mean would be smaller than the median.

Positive Skewed:
The mean is pulled in the direction of the high scores (the tail to the right). When a distribution is positively skewed, the mean is larger than the median.

25
Q

Chi2

A

is used with NOMIAL data and compares observed frequencies with expected frequencies. It may be used when a study has only one group of subjects.

26
Q

T-test

A

is used to determine whether there is a statistical significance between the means of two groups. It is used with INTERVAL AND RATIO scale data and when the researcher has a treatment group(s) and a control group to see if the treatment makes a difference.

27
Q

ANOVA

A

is like multiple T-tests and is used with three or more

groups.

28
Q

MANOVA

A

Shows the correlation between each independent variable and the dependent variable (because multiple
independent variables are used).

29
Q

ANCOVA

A

Shows how a covariate interacts (co-varies) with the dependent variable. A covariate is a variable correlated
with the dependent variable.

30
Q

Types of Error

Type I

A

A null hypothesis is rejected when no differences exist. Remember, a null hypothesis is stated, “There is no significant difference……” In a Type I or Alpha
Error, the researcher says there is a significant difference (thereby rejecting the null hypothesis) when there really isn’t a significant difference. In other words, the researcher
rejects the null hypothesis as originally stated when he/she should have accepted it.

31
Q

Types of Error

Type 2

A

A null hypothesis is accepted or retained when differences do exist. Remember, a null hypothesis is stated, “There is no significant difference….” In a Type II or Beta Error, the researcher says there is no (significant) difference (thereby accepting the null hypothesis) when there really is a significant difference. In other words, the researcher
accepts the null hypothesis as originally stated when he/she should have rejected it.